Writing Accross Boarders

The Idea of Writing

Writing Across Borders

Edited by

Alex de Voogt and J oachim Friedrich Quack

BRILL

LEIDEN . BOSTON 2012

Cover iLLustration: Multi-lingual sign posted by the Greek-Norwegian mission on Sai Island, Sudan, using Arabic and Old Nubian scripts for the Nubian language.

This book is printed on acid-free paper.

Library of Congress Cataloging-in-Publication Data

The idea of writing : writing across borders / edited by Alex de Voogt and Joachim Friedrich Quack.

p. cm. Includes index. ISBN 978-90-04-21545-0 (alk. paper) 1. Writing. 2. Written communication. 3. Language and languages-Orthography and spelling.

I. Voogt, Alexander J. de. 11. Quack, Joachim Friedrich, 1966-P211.134 2011 302.2'244-dC23

This publication has been typeset in the multilingual "Brill" typeface. With over 5,100 characters covering Latin, IPA, Greek, and Cyrillic, this typeface is especially suitable for use in the humanities. For more information, please see www.brill.nl/brill-typeface.

ISBN 978 90 04 21545 0 (hardback) ISBN 978 90 04 21700 3 (e-book)

Copyright 2012 by Koninklijke Brill NV, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill, Global Oriental, Hotei Publishing, mc Publishers, Martinus Nijhoff Publishers and VSP.

All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher.

Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA Fees are subject to change.

r'\ " -u-> Paper from responsible sources

!.�� FSce C008919

MIX

PRINTED BY AD DRUK BV - ZEIST, THE NETHERLANDS

CONTENTS

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . vii

Invention and Borrowing in the Development and Dispersal of Writing Systems ........................... . .............................................................. 1 Alex de Voogt

27-30-22-26 - How Many Letters Needs an Alphabet? The Case of Semitic ...................................................................................................... 11 Reinhard C. Lehmann

Nubian Graffiti Messages and the History of Writing in the Sudanese Nile Basin ............................ . . .................................................... 53 Alex de Voogt & Hans-Jorg Dohla

About "Short" Names of Letters Konstantin Pozdniakov

Early Adaptations of the Korean Script to Render Foreign

69

Languages ...................................... .......................................... .. . ......... ......... 83 Sven Osterkamp

Han'gill Reform Movement in the Twentieth Century: Roman Pressure on Korean Writing ................................................................... 103 Thorsten Traulsen

The Character of the Indian Kharo�thl Script and the "Sanskrit Revolution": A Writing System Between Identity and Assimilation .......................................... . . ....... .............................................. 131 Ingo Strauch

Symmetry and Asymmetry, Chinese Writing in Japan: The Case of Kojiki (712) . . ... . . . . . . . . . . .. . . ... . . . . . . . . . .... . . . . . . . . . . ... . . . . . . . .... . . . . . .. . ... . . . ... . . . . . . . . . . . . . . . . . . 169 Aldo Tollini

vi CONTENTS

Writing Semitic with Cuneiform Script. The Interaction of Sumerian and Akkadian Orthography in the Second Half of the Third Millennium BC ..... . ... ......... . .. . .... . . . .. ............. . ... ..... . . ................ . 181

Theo JH. Krispijn

Old Wine in New Wineskins? How to Write Classical Egyptian Rituals in More Modern Writing Systems .. ................................... ... . . 219 Joachim Quack

Subject Index .................................................................................................... 245 Language (Group) and Script Index ............. ....... ..... ........... .. ....... . .... . ...... 249 Author Index ....... .................................... .... ............. ................ ........... ... . . .... .... 251

ACKNOWLEDGEMENTS

The Cluster of Excellence "Asia and Europe in a Global Context: Shifting Asymmetries in Cultural Flows" at Heidelberg University hosted the exchange of thoughts and ideas that were at the basis of this publication. In particular, we wish to thank Profs. Madeleine Herren-Oesch, Axel Michaels and Rudolf Wagner for their support.

The Idea of Writing seminar series started at Leiden University where it was hosted for five years. This is the first volume that resulted from a meeting outside Leiden, at Heidelberg University in Germany. Subsequent seminars have been held in Einsiedeln, near Zurich, Switzerland and Venice, Italy. They illustrate the lasting interest that this exchange has created. We owe particular thanks again to Connie Dickmeyer for her corrections and suggestions, and the patient staff at Brill Publications who have facilitated the continued productivity of these exchanges by publishing this second volume of contributions.

The publication of this volume has been made possible by generous funding of the Deutsche Forschungsgemeinschaft (German Research Foundation) in the context of the cluster of Excellence "Asia and Europe in a Global Context".

27-30-22-26 - HOW MANY LETTERS NEEDS AN ALPHABET? THE CASE OF SEMITIC*

Reinhard G. Lehmann

Introduction and General Considerations

In his recent novel "Das Geheimnis des Kalligraphen", the Syro-German writer Rafik Schami reports a fictitious dialogue between the progressive young calligrapher Hamid Farsi and his master Serani. In the course of this dialogue, the master states:

Die arabische Sprache hat nur neunundzwanzig Buchstaben, und je mehr du davon zerstOrst, umso unsicherer und ungenauer wird die Sprache. (Schami 2008: 395)'

Modern oriental scholars know that the Arabic alphabet has 28 letters. A twenty-ninth letter is added by tradition, attributed to Mohammed himself, and is nothing but the Lam-Alif ligature LA, which simply combines the two letters into a single cohesive unit. Accordingly, Hamid Farsi considers the Lam-Alif ligature superfluous, artificial, and not a genuine letter in the Arabic alphabet. However, the more conservative and careful calligraphy master Serani warns his young, impatient pupil not to be all too optimistic, modernistic, or straightforward:

Ich will dich nicht entmutigen. Diesen Buchstaben LA hat der Prophet dem Alphabet hinzugefiigt, und er bleibt, bis die Erde untergeht. Wenn du auf mich horen willst, streiche keinen einzigen Buchstaben, weil sonst die ganze islamische Welt gegen dich sein wird, denn diese Buchstaben kommen im Koran vor. (Schami 2008: 39SY

* Thanks to Robert M. Kerr, Philip C. Schmitz and other colleagues and friends with whom I could discuss parts of this paper. However, mistakes and errors are exclUSively my responsibility.

, The Arabic language has only twenty-nine letters, and the more of them you destroy, the more uncertain and imprecise the language becomes. (translation by R.G.L. ) .

2 I don't want to discourage you. The Prophet himself added the letter LA to the alphabet, and it will persist until the world ends. If you listen to me, then you won't do away with one single letter, or otherwise the whole of the Islamic world will disapprove of you because these letters are found in the Koran. (translation by R.G.L.) .

12 REINHARD G. LEHMANN

And:

Man nimmt keinen Buchstaben weg [ . . . ] den die Jahrtausende geformt haben . . . (Schami 2008: 396)3

The dialogue scene points to some problems, which are directly relevant to the title of this paper:

- To determine how many letters, or graphemes, an alphabet needs, it i s crucial to know what a letter is. Is the Lam-Alif ligature of the Arabic alphabet a letter, or is it not? Moreover, ifLam-Alif is not a letter, why then should OMEGA, the last letter of the Greek alphabet, be labelled a letter? Yet, it is conspicuous that even the name O-mega tells us that it is nothing more than a diacritical variation to denote the 'big'-that is, long-pronunciation of the {o} of the Greek alphabet, which thus became the O-mikron. This is a fact that is also proven occasionally by its early archaic written forms.4 And if the Arabic LA were a letter, why then isn't the German digraph {CH} or the German {SZ} ligature (g) a letter?

- There is a claim that reducing the number of letters of an alphabet causes uncertainty and imponderability of the language. Is this claim really convincing?

- There is the claim that a letter, once invented, will last for eternity: "You must not remove a single character [ . . . ] , which has been formed by millennia. " This, however, is obviously not true. In the course of his tory, several letters have become obsolete and were removed in several alphabets.

Despite the traditional Muslim's claim that the Arabic alphabet has 29 letters, modern oriental scholars know that it actually has only 28 letters. And some might even say that the Arabic alphabet has only 15 or

3 You must not remove a single character [ . . . ], which has been formed by millennia. (translation by R.G.L.) .

4 In areas like Ionia, Knidos, Paros, Melos and others, the difference between the short and long /0/ was marked by use of "a new sign which appears to be a doublet formed from 0 by breaking the circle" Qeffery 1961: 37) , i.e., opening it at the bottom. Another interesting specimen is found in an archaic inscription from Phleious (first half of the sixth century BeE) where in the word AWL the OMEGA is written with a digraph formed by two full circle dotted OMIKRONs that are written one below the other Qeffery 1961: pI. 24,lb, also Jeffery 1961: 147 note 1). On this basis, on the other hand, Bernal even concludes that the double-circle was its most pristine shape (BernaI 1987: IS; 1990 : 121).

27-30-22-26 - HOW MANY LETTERS NEEDS AN ALPHABET? 13

18 letters, because some similar letters have merged graphically, such as ra with zay, /:la with fJa and gim, ha with ta and fa (and, the final shape aside, also with ya and nun) , and fa with qaf, or were secondarily differentiated for phonemic reasons, such as dal and gal, and' ayn and gayn, and were thenceforward distinguishable only by additive points (Gruendler 2006: 151) . These distinguishing points, or diacritic marks, have not been there from the beginning of the Arabic script, and are not in the earliest versions of the Qur'an (Gruendler 1993: 131SS) . But others may argue that in earlier times, when the Arabic script evolved from an older, most presumably Nabataean or Nabataean-like ancestor, these letters of course were distinguishable, and only later merged into one common shape, thus making it necessary to add diacritical marks, or points (Gruendler 2006) . This, however, implies that older Arabic or pre-Arabic texts for instance remained readable for some time anyway, even after letter shapes became very similar or merged totally (Kerr 2010: 372) . It also implies that neither the merger or reduction of a graphemic inventory nor its enhancement by diacritics had its most important reason in simple readability.

Thus, how many letters does the Arabic alphabet, for instance, have? And how many letters does an alphabet need at all? Anyway, the quest for the number of letters that form an alphabet could indeed be a bit complicated.

The International Phonetic Alphabet (IPA) contains more than 130 graphemes, or signs, and depending on what you really count, there are even more than 160 signs, or 'letters'. However, no one would reasonably say that the IPA is a real natural alphabet. There is presumably no individual that will actually be able to produce in speech all the sounds of this set of signs, or feels the need to do so , simply because there is not a

Single language that has all of these phonemes. Thus, there is a remarkable restriction: an alphabet represents only the sounds of a single natural language (or group of natural languages ), which is used by a certain group of speakers at a certain time in history. Does it represent all of the sounds of such a natural language? Presumably not-there is sufficient evidence to assume that every natural alphabet as a graphemic system is to a certain extent deficient in relation to the phonemic system of the language or dialect it is used for. Moreover, obviously the ratio between a phonemic and a graphemic system can change in the course of history.

I will suggest in this paper that the Northwest Semitic shortened or 'short Abgad' does not reflect a spoken dialect (the Phoenician, as is mostly believed) at all, but has emerged as a scriptio franca for the Semiticspeaking Levant. In the late second millennium BeE, the short Abgad,


which eventually made the game in alphabet history, provided sufficient consonantal contrast to constitute a to some extent supraphonemic 'writing across the borders' that covers a broad variety of Levantine Northwest Semitic dialects. To get this point, a Semitic alphabet typology, comments on several questions and unsolved problems, and a critical evaluation of terminology are required.

Historical (and Typological) Background

There is no doubt that what we call an alphabet today had developed more than 3000 years ago somewhere in Syro-Palestine, or the Levant. However, a closer look at these ancient Levantine alphabetic ideas shows conspicuously that already in antiquity there were two competing alphabetical sequences: Hala/:tama and Abgad.5 Both seem most probably to have displayed close and presumably almost full match of both, the existing phoneme inventory of the language and the provided graphemes, and, what is more, both seem to have originally reflected a common phonetic inventory, albeit in a different order. Of course, because they did not represent vowels, some would argue that they both are not true alphabets. Actually they really were. We will come back to this point again later.

Old South Arabic - Sabaean (Sabaic): 29

What has been traditionally called the 'South Arabic' sequence, or alphabetic order, also called Hala/:tama according to its first four letters h-l-/:t-m, is almost exclusively attested in the South-Arabic language area. From several finds of mainly Sabaic abecedaries, this fixed alphabetic Hala/:tama letter sequence is known already from the first millennium BCE as follows:

h-l-/:t-m-q-w-s-r-b-t -s-k-n-b-$-s-f-' -' -r;l-g-d-g-t-z-g-y-t-�

Minor deviations from this sequence are only found in some stone inscriptions, but not so far in the so-called minuscule script, which was used on

5 Of course, there are also other 'alphabetic' sequences of other scripts in antiquitiy, the most interesting of which is that one used by the older runic script, called the Futhark (fujJark). Theo Vennemann in several publications recently tried to derive it from the Phoenician-Punic Abgad-sequence, which in my opinion is the most convincing theory up to now (Vennemann 2006, 2009).

wooden sticks, esp. palm-leaf stalks.6 With 29 phonemes, the old South Arabic alphabet seems to preserve most completely the original Semitic consonantal phonemic inventory (Stein 2003: 18) .7

Based on palaeographical criteria, the Old South Arabic script would seem to date to at least the mid-second millennium BCE. This presumably also applies to the Halal:tama sequence. This script was continued independently in Ethiopia where a South-Semitic language (Ge'ez) was adopted early in the Common era. The Ethiopic derivative of the 'South Semitic' writing system in a separate and isolated development due to phonetic mergers often confuses some graphemes in writing. It was also secondarily transformed into a 'semisyllabary' (Diringer 1968) by altering the shapes of the consonantal graphemes depending on the following vowel. Hence, one can claim that the classical Ethiopic script has either 26' consonantal graphemes or is a syllabary with 251 characters (33 x 7 +

20; cf. Getatchew Haile 1996: 573). Nonetheless, the Ethiopic script still "has the analytic depth of an alphabet" ( Coulmas 1989: 153) .

However, despite the almost exclusive South Arabian attestation of the Hala/:tama sequence and its seemingly exclusive later South Semitic history, it is noteworthy that the very earliest evidence of Hala/:tama is found in the thirteenth century BCE long alphabet of 27 signs from Ugarit in Syria ( see below).8 Accordingly, the so-called 'South Semitic' alphabet is

6 Stein (2003: 11, 2010: 591); for further references Stein (2010: 591 and 743f) . Abecedaries written in stone are published by: Bron & Robin (1974), Irvine & Beeston (1988), Ryckmans (1985). The sole published deviation in minuscule script, with the interchange of d and z in Oost.Inst. 37, most probably is nothing but a (mistaken) exception (Stein 2003: 11 Fn. 64).

7 A remarkable phenomenon regarding the number of letters is that only in minuscule writing, i.e. on wooden sticks, the South Arabic alphabet uses the grapheme <cj>, where for etymological reasons <f:> is to be expected. But in terms of lexicography, there are no exceptions (Stein 2003: 27f, Nebes & Stein 2008, and recently Stein 2010: 502f) .

8 A reverberation of the HalaJ:tama is also found in a late Egyptian alphabetic device, as recently is shown by Quack 1993, 2003. Moreover there is the claim that an underlying notion of HalaJ:tama were also perceivable in the Latin word elementum ("letter / character as, basic constituent of speech") , reflecting ha-la-/:la-ma (Muller 1994: 309, Gruendler 2006: 148) . At any rate, this would mean that the so far only rarely attested northern, or Levantine, history of the Hala/:lama should not have come to an end so early after all. However, though for phonological reasons it is most probable that Latin elementum is a foreign loan word ("Herkunft unklar, jedenfalls nicht Erbwort", Walde & Hofmann 1938: 398), MUller's assumption has severe historical difficulties. If Latin elementum depends on ha-la-/:la-ma, it would be the only first millennium BeE remnant of the Hala/:lama in the Mediterranean except for the above-mentioned Egyptian and a few Ancient North Arabic 'abecedary' inscriptions in different, if any, order including a modified Abgad as well as Hala/:lama (Macdonald 1986, 2008a: 185) , which then must have been responsible for the transmission onto Latin. Yet, also the older hypothesis that elementum represents

REINHARD G. LEHMANN

by no means exclusively South, but also Levantine! This is not even ruled out by an occasionally reported assumption that the Ugaritic HalaJ:zama were only the attempt of an Arabian tourist scribe to learn the Ugaritic script-which is nothing but an over-romanticized attempt by modern scholars to escape the aporia of the facts.

The Ugaritic Evidence: 30 (27)

It is a matter of fact that the earliest well attested rich inventory alphabetic system with a considerable number of texts known hitherto is the Ugaritic cuneiform, or wedge alphabet. The texts and abecedaries that have been found in Ugarit and its environs witness to three different systems of a cuneiform alphabet.

Currently, eighteen Ugaritic cuneiform abecedaries, combined textabecedaries or abecedary fragments are known (Puech 1986, Hawley 2008) . Nearly all of these abecedaries represent the dextrograde Ugaritic long alphabet of 30 signs. Its original 27 signs were most probably adapted from an unattested Northwest Semitic linear alphabet tradition (Lundin 1987b, Dietrich & Loretz 1988) . Three additional signs were doubtlessly secondarily appended hence finally prodUcing a 30-letter sequence. As nearly all of these abecedaries9 show, they baSically arrange the characters according to an expanded scale of the so-called Abgad device, which has its name from the first four letters' -b-g-d of the short standard scale (see below) :

'-b-g-fJ-d-h-w-z-J:z-t-y-k-s-l-m-d-n-f,-s-' -p-$-q-r-J-g-t-' i-'u-s

the beginning of the second half of the Abgad order (el-em-en- . . . ) in analogy to 'Alphabet' (Walde & Hofmann 1938: 398) fails because such half-cutting division of the alphabet is not attested in Latin, but in Semitic only (which can be seen by the Atbash encoding device, or, for instance, Psalms 9 and 10, which subdivide a totally complete alphabetic acrostic in its middle, and by Nahum 1, which covers only the first half of the alphabet. See for a recent discussion Renz 2009). Moreover, the pronunciation of the letters L, M and N as 'el', 'em', and 'en' seems to be not much earlier than the fourth century CE. Thus, there seems to be an interdisciplinary circular reasoning, and regarding the etymology of elementum Latin philology seems to be as helpless as it was one hundred years before (thanks to Christian Tomau, Wiirzburg, with whom I was able to discuss this pOint).

9 These abecedaries are listed in Hawley 2008 and in KTCP 5.4-5.6, 5.9, 5.12-5.14, 5.16-5.17, 5.19-5.21, 5.30. Only in the fragments RS 15.071 (KTCP 5.8), 5.274 (7.54), 12.019 (5.5), and 19.174[4] (9.342) is no decision possible whether they contained 30 or only 27 letters. The fragmentary abecedary RS 16.265 covers the missing letters in additional enigmatic (exercise ) letter sequences that remind of a similar phenomenon in lines 1-4 of the proto-Hebrew 22-letter abecedary ostracon from 'Izbet $artah. For RS 88.2218 (dextrograde Halaf:tama sequence) , see below. For an overview and discussion of Ugaritic abecedaries see Pardee (2007).

This first category, the cuneiform long alphabet, is by far the most widely used, represented in numerous administrative and literary texts from the north Syrian coastal sites of Ugarit and Ras Ibn HanL Noteworthy, however, is the occasional interchange of signs of mostly similar phonemic value within the Ugaritic Long Alphabet, thus <t> for <$>, as in art for ar$ (affricate pronounciation), <w> for <h>, as in bwtm for bhtm, <h> for <fJ> as in phr instead of pfJr, and < ' > for < ' > as ttar instead of tfr (Tropper 2000: 21.352.5) . Maybe such examples are more than mistakes only, and reflect vernacular variations that might have rendered similar phonemes with a

reduced set of graphemes. Another long but non-expanded cuneiform alphabet of 27 signs is also

attested at Ugarit, but only in one abecedary that is written dextrograde (RS 88.2215 = KTU 9.426, Bordreuil & Pardee 1995, R6llig 1998, Bordreuil & Pardee 2001) . It omits at least the last three letters of the Ugaritic Standard long Alphabet and displays some more peculiarities in both letter shape and adjustment. A similar abecedary was discovered in Beth Shemesh to the west of Jerusalem, but is written sinistrograde (KTU 5.24 = 8.1, Lundin 1987, Pardee 2003/04: 18 n. 43) . Both, though not entirely identical, bear witness to the so-called South Semitic alphabetic Hala/:lama device. lO However, to date no Ugaritic texts are known, that are unambiguously written in this 27-letter alphabetic system.

Finally, at Ugarit there was a possibly somewhat later cuneiform short alphabet system of presumably 22 letters, which could be written in either direction (Dietrich & Loretz 1988: 145-275) .11 Texts written in this alphabetical system seem to be dated not earlier than the thirteenth century BeE, and are not only found three or four times in Ugarit/2 However, they also spread south from Ugarit across the Levant, to several sites

10 "Enfin, les formes des signes de cet abecedaire si particulier sont tout aussi etrangeres a Ougarit que l'ordre: aucun autre texte ougaritique n'utilise cet inventaire graphique" (Hawley 2008: 225 ) , similarly Robin (2008: 233) . However, Rollig (1998: 87 ) states "An ugaritischen Schreibgewohnheiten andert dieses Alphabet ja nichts, denn der Phonembestand blieb derselbe. Man wird also in der ugaritischen Uberliefemng vergeblich nach Texten im 'siidsemitschen' Alphabet such en." According to Rollig, the few differing sign forms could be blamed on a scribal scholar and are not evidence enough for use or non-use of the Hala/:lama series in Ugarit.

11 As yet, there are only 21 letters of the Abgad device identified, missing the <$> sign. Also possible is that there was a short alphabet tradition with 24 letters, missing only the interdentals of the long alphabet (Tropper 2000: 76-77 ) . A table of the sign forms, which vary to a certain extend, is found in Tropper (2000: 75) .

12 Ugarit: KTU 4.31 (right to left), KTU 4.710, and Minet el-Beida: KTU 1.77 (right to left). The attribution of a further text, the votive clay nail KTU 7.60 to this group is questionable, Dietrich & Loretz (1988: 168-70; 1989: 107) .

REINHARD G. LEHMANN

in Syria-Palestine and Cyprus/3 Some of them might also have 24 letters, what could be counted as a transitional stage from the Long to the Short Alphabet. 14 Nevertheless, they bear witness to a short( ened) alphabetic cuneiform tradition. Often this short cuneiform alphabet is considered the day adaptation of the Phoenician linear alphabet into wedge script (Pardee 2008: 7), which implies that the Phoenician alphabet was an entirely independent development (however, the question remains why there should have been such radical reduction in the phoneme inventory only some 200 km south in 'Phoenicia' ) . In this Ugaritic short alphabet, there is only one letter for ale! (Le., without vowel-connotation), the <!J> sign is replaced by the </:l > sign, the <[j> sign is replaced by the <

' > sign

(which indicates a merger of the two sounds-or graphemes-respectively) . Also the interdental phonemes It/, I�/ and Idl have no distinctive grapheme but seem to be represented by their sibilant equivalents, i .e . , /s/, 1$1, and /z/ respectively (Dietrich & Loretz 1989: 107f, Tropper 2000: 22.5) . The 'short' alphabet was used to write texts; no abecedary is yet attested in this alphabet, and even the few examples of these texts found at Ugarit were all from outside the palace area, thus suggesting that they do not originate from the 'official' schools ofUgarit (Dietrich & Loretz 1989: 108) . Hence, we do not know exactly whether the Ugaritic short alphabet bears witness to the Abgad sequence (though this is not unlikely), and it is uncertain whether the texts constitute a 'phoenicianizing' representation of the Ugaritic language or are in a kind of early plain Phoenician or another North Canaanite dialect (Pardee 2003/04:17). Anyway, lexicon and grammar of the short alphabetic cuneiform texts seem to be closer to later Phoenician or another North Canaanite dialect than to the Ugaritic texts of the long alphabet (Tropper 2000: 22.81ff; Dietrich & Loretz 1989: 110) .

With the destruction of Ugarit at the beginning of the twelfth century BCE the Ugaritic cuneiform alphabet tradition came to a sudden end. But

13 Listed, for instance, by Millard (2007: 85), which are: Silver Bowl from Rala Sultan Tekke (Cyprus) 6.68; storage jar from Tell Nebi Mend/Qadesh (but with the <FP sign, thus maybe long alphabet tradition); two vessel inscriptions from Kamid el Lodz (one dextrograde, the other sinistrograde); an inscribed jar handle from Sarafand-Sarepta [Lebanon] KTU 6.70 (dextrograde, relative z and verb p'l indicating 'Phoenician' Canaanite language : Smith (2005) ; Vita (2003: 401f) ; a bronze tablet (knife blade?) from Tabor valley (dextrograde) KTU 6.1; a small tablet from Tel Taanak (left-to-right) KTU 4.767. Dietrich & Loretz (1989: 109 ) : "This group of texts attests inscriptions written in both directions. The widely held notion, that the short alphabet is characterized by right-to-Ieft writing cannot be maintained. In these texts the direction of writing appears to be rather a function of local school traditions."

14 KTU 6.68 and KTU 5.24 including the <Ip and <FP sign (Tropper 2000: 22-45 ) .

the Abgad order that formed its dominant arrangement device (but most likely was not an Ugaritic invention, nor was alphabetic writing at all, Pardee 2007: 182) survived as the basic order of most Western and Semitic alphabets until today.

The arrangement of the short alphabet in a 22-letter Abgad device as it is attested from the late second millennium BCE onwards seems to be nothing but an abbreviation, or selection from the (Ugaritic) long Abgad series. Its most common standard device, now starting plainly with '-bg-d, is known from several Northwest Semitic abecedaries of the first millennium BCE, but best from the Hebrew Bible:'5

'-b-g-d-h-w-z-/:z-t-y-k-l-m-n-s-'-p-$-q-r-s-t

The as yet oldest authority for a 22-letter short Abgad sequence, however, is the proto-Hebrew linear exercise on the 'Izbet $artah sherd, datedby archaeological context data-to the late twelfth century. In its bottom line, it clearly displays a genuine Abgad device, although with slight alterations (Cross 1980; Hallo 2004: 28Sf):

'-b-g-d-h-m-l).-z-t-y-k-l-[ ]-n-s-p-'-$-q-q-s-t

A quite similar arrangement is found in the late 10th century Tel Zayit stone abecedary (Tappy et al. 2006; Sanders 2008; Rollston 2008b) , which was discovered in 2005:

'-b-g-d-w-h-l).-z-t-y-I-k-m-n-s-p- '-$-q-r-s-t

While the apparent position change of the bilabial sounds /w/ and /m/ in the 'Izbet $artah sherd most probably had phonemic reasons, the </:z>-<z> and the <' >- interchanges here (and in a few other abecedaries) are remarkable.

The latter is also found in some other proto-Hebrew (and Hebrewonly, or Southern Levantine, but not Ugaritic ! ) abecedaries of the early

15 Unfortunately, the numerous Northwest Semitic linear abecedaries are mostly either fragmentary or represent a selection of some letters only. However, these all incontrovertibly bear witness to the Abgad series or at least segments thereof. In the Hebrew Bible, though, there is a considerable number of complete abecedaries attested in the follOwing Biblical Hebrew acrostics, where every colon, verse, or stanza starts with a new letter of the alphabet: Psalms 9-10 (disturbed between Lamed and Pe) , 25 (Waw missing), 34 (Waw missing) , 37, Ill, 112, 119, 145 (Nun missing) , Prov 31:10-31, Lamentations 1, 2, 3, 4, and Sirach 51:13-30 (Zayin-Kaf missing), Nah 1:2-8 (first half), outside the Bible also in the Dead Sea Psalms scroll 11QpSa XXi:ll-17, xxii:I-15, xxiv:3-17. For Biblical Hebrew acrostics in general, see Freedman (1999 : 1-24) , for abecedaries, most recently Gzella (forthcoming).

first millennium, the most prominent of which are the abecedaries 1 and 3 from Kuntillet Adjrud in the southern Negev.16 Moreover, this very peculiarity of an < '>- to -< '> interchange is also found in some Biblical acrostic poems.17 Though, in any case, this deviation from the majority or 'standard' sequence does not have a phonemic reason, it is not likely that it was just haphazard or mistaken. Rather, its long afterlife even in the Hebrew Bible seems to trace back to a different Abgad variety, which must be seen in connection with local or areal 'Hebrew' scribal traditions.

Thus, to conclude the first overview, there seem to have been five main classes of early alphabet records, to be categorized by graphic format (cuneiform or linear), number of graphemes, abecedary-device, and direction of writing:

,6 This -< '> modified Abgad is also found 3 times on an unprovenanced early sixth century BCE Hebrew ostracon of the Moussaieff collection, which was recently published by Martin Heide (Heide 2007).

' 7 Among the aforementioned (n. 15) , a modified order with interchange of < '>- to -<'> is found in Psalms *9-*10 and Lamentations 2, 3 and 4, and also in the assumed underlying Hebrew text of the Greek (Septuagint) version of Prov. 31. That this -< '> modified Abgad device is as yet found outside the Hebrew Bible in South Levantine or 'proto'-Hebrew abecedaries only, is maybe of some importance. Anyway, the majority of later Abgad abecedaries in general bears witness to the < '>- order, which accordingly we may label the 'standard'. It is not only found in most Biblical poetic acrostics (see note 15), but in an early ostracon of the eighth century from Lachish, too. Unfortunately though, most Hebrew Abgad abecedary inscriptions and also some Phoenician, Aramaic, and Ammonite abecedary sherds or seals are either fragmentary or too short respectively to reveal whether they testifY to a < '>- or to a -< ') order. Accordingly, nothing can be said about which of both variants was the most common. Quite differently, Sanders (200S: 102, referring to Byrne 2007: 4-5) claimed the -< '> sequence to be the older one because of the well known Hebrew scribal habit of placing a directly under a 'superscript' <s>, which, in his opinion, "could well have been triggered" as a "ghostly trace of a different order entirely" by the Halaf.zama device with its <s>--< '>-< '> sequence. However, Byrne's underlying notion is not convincing. Abecedaries, whatever their main function might have been, were not very useful as a curricular guide (as Byrne 2007: 5 claims) "to gain familiarity and proficiency with the characters, their relative size, and their relative placement" (Byrne 2007: 5). Would it not have been much more effective to gain scribal proficiency in baseline control, relative letter size, and letter placement by copying texts (not abecedaries), for these alone provide most or all of the combining possibilities? Beyond sole reading and writing knowledge, mastership-level scribal proficiency is first and foremost knowledge of layout and keming devices (Lehmann 200Sb), which were acquired by practicing, writing and copying texts again and again. Moreover, Sanders obviously ignores the fact that the -< '> sequence, which he claims to be "found in no Israelite abecedary", is well attested at least three times in the Hebrew Bible in reasonably late texts (Lamentations 2-4), which alone rules out any idea that this order might be the earlier, Canaanite, and non-Israelite.

Cuneiform (wedge) long alphabet with expanded inventory (30 ) : Abgad device, left-to-right (dextrograde) Cuneiform (wedge) alphabet with full inventory (27) : Hala/:tama device, dextrograde and sinistrograde ( so far only two examples) Cuneiform (wedge) alphabet with reduced inventory (22 ) : Abgad device? (no abecedary), sinistrograde (right-to-Ieft) Linear alphabet with full inventory (27 ) : Abgad device? (no abecedary), sinistrograde Linear alphabet with reduced inventory (22) : Abgad device, sinistrograde

To be honest-there is a 'fake', or misleading assumption, in this overview because there actually is no full-inventory 27-sign linear alphabet as yet found. This is only what scholars normally expect to have been there.Is Nevertheless it is not attested. Maybe there are some non-linear, quasipictographic forerunners of the rich-inventory alphabet in the early second millennium, depending on how you interpret the protosinaitic and related inscriptions of the second millennium BCE. Nonetheless, they are not linear. As yet, the full inventory 27-sign linear alphabet is a phantom, as is a yet-unattested 27-sign cuneiform (wedge) Abgad sequence. However, while there are good reasons to assume the existence of the latter/9

18 See, for instance, the entirely useless effort made by Martin (1962) to identify in the sub-text of the so-called Rapa' palimpsest arrowhead a linear <g>-letter, bearing the shape of <z> "with two small slanting strokes appended to the upper crossbar" (p. 183) . All necessary obj ections to this is said by Cross (1967: 14* n. 34). Relying strictly on the claim already made by W.F. Albright that there must have been a Northwest-Semitic linear long 27-letter-alphabet, as far as I know, Martin was the last who seriously tried to find something like that in any Northwest Semitic linear alphabetic inscription. It is obvious that Martin failed totally. As a side-effect of his fatal methodological meander, unfortunately also other useful and valuable insights of Martin remained underestimated.

19 The final position of the surplus letters 'i, 'u, s in the expanded cuneiform Abgad points clearly to a primal 27-letter Abgad that closed with the letter <t>o Originally, there was one '-sign only, which was a pure consonantal glottal stop (the 'a sign) . The vowelspecializing signs 'i and 'u, as well as s predominantly for loanwords, were added at a later stage (Sivan 2001: 9-10) . Furthermore, the abecedary RS 230492 ( = KTU 5.19) has an unexpected and otherwise inexplicable dividing stroke after the <t>-sign, which confirms that the users of the 30-letter cuneiform Abgad by all means were aware of the secondary character of the last three signs 'I, 'u and s (Bordreuil 1982: 9-10: " . . . confirme que les utilisateurs de l'alphabet ougaritique de trente lettres etaient conscients du charactere supplementaire des trois dernieres lettres i, u, et s."). However, the KTU edition as well as Tropper (2000: 22) interpret the divider as a missing part of the 'i sign, which, from a palaeographic point of view, is less probable.


it remains unlikely that there ever was something like a linear long Abgad. Thus, the fourth class in the above list should be disregarded.

As for the remaining classes, there are two basic parameters: the arrangement of letters, and the number of graphemes. It seems as if here lies the historical clue for the question under discussion. Beforehand, however, a clarification of terminology is necessary.

Abjad and Alphabeta-the Vaniels'Distinction

Among the four abovementioned parameters, graphic format and direction-of-writing are predominantly dependent on extrinsic factors such as areal economic conditions and availability of writing material, or sociological and regional trends of scribal behaviour. 20 Thus, apparently it seems most important to keep apart the number of graphemes and the abecedary-device, which are the basic classes that point more to the intrinsic notion of writing itself. But not all scholars do so.

In the last two decades, a traditional term for arrangement, or abecedary order, became almost totally intermingled with what grammatologists hold important in terms of 'alphabet' type-with fatal effects on terminology. Maybe some would object that the abecedary order is not relevant. But the chosen term suggests that the order of signs is relevantwhich, indeed, it is. Sadly, this is how terms got confused.

It was Peter T. Daniels who brought up a terminological opposition between Alphabet and Abjad. He first did so in 1990 to refute I.J. Gelb's earlier statement that West Semitic scripts were not alphabets, but rather syllabaries (Daniels 1990, Gelb 1963, see also Powell 1991: 238-245 and Powell 2009: 153 ) . Meanwhile, this distinction has attained certain notoriety amongst linguists and, above all, on many Internet sites. Respond-

20 It is clear that the Ugaritic cuneiform alphabetic writing was influenced qua technique (extrinsic factors) by the Hittito-Mesopotamian writing traditions. Nevertheless, it shares its basic intrinsic principle (alphabet) with the bulk of Semitic linear writing. On the other hand, a close investigation of the Northwest Semitic script traditions of the first half of the first millennium BeE has shown that script changes up to seemingly different 'national scripts' depend almost exclUSively on changes in the scribe's hand and the scribe's attitude, that is: on regional sweep and influence of dominant scribal schools (van der Kooij 1986: 90-93.244-251.253) . As long as there is no better explanation, we must assume that also the earlier change and fixing of writing direction( s) had the same cause. Once a dominant writing direction had been set (however marginal the reason might have been), the canon of 'path dependence' became the rationale for all further development and did not allow any reconversion of direction or shape in writing.


ing to an earlier objection to Gelb by Swiggers 1984, who differentiated between a 'standing-for' and a 'denotation' view, and because "distinct types of script should not be lumped together" (Daniels 1990: 729 ) , Daniels wanted to further differentiate between terms than just 'syllabary' and 'alphabet' alone. This is why to define Alphabet more rigorously he introduces the traditional Arabic numeric letter sequence Abjad as a new script term into grammatology, and in the subsequent years he has vigorously defended it. An alphabet, Daniels claims, "contains characters that denote all or most of the individual segments of a language, both vocalic and consonantaL . . . The West Semitic scripts, then, are not alphabets, since they do not represent the vowels." Rather, the West Semitic scripts "constitute a third fundamental type of script, the kind that denotes individual consonants only" (Daniels 1990: 729) , for which he proposes to use the traditional Arabic numerological term Abjad.

Only the Greeks, Daniels claimed, first developed what one could label a 'true' alphabet. However, he did not say so because of the etymology of the word (which would be a true and irrefutable reason) , but because a 'true' alphabet in his view must represent all or nearly all phonemes of a language including vowels-which obviously is not the case with any ancient West Semitic 'alphabet'. To escape objections regarding the vowelled aleph-signs in the Ugaritic expanded 3o-letter Abjad for instance, he simply labels this writing system an 'augmented Abjad' (Daniels 1990: 730 ) . By the way, he even introduced a "fourth fundamental type of script," though not as new, when he along the way redefines the older term 'neosyllabary' for the Ethiopic script system and labels it Abugida (Daniels 1990: 730) .2 1

It is fully comprehenSible that in his zeal to refute the Gelb thesis, Daniels creates another and more sophisticated system of his own. Some years later only, in astounding aplomb he then stated: "It must simply be recognized [ . . . ] that abjads are not (any longer) syllabaries and not (yet) alphabets, and that abugidas-though they denote syllables-are not like syllabaries, since vowels receive identification equivalent to that for consonants." (Daniels 1996a: 8 ) . Notwithstanding the fact that the apparently sophisticated nomenclature used by Daniels has already found wide

21 It is interesting that Daniels here is still cautious: 'Were it not for the existing term, I would propose maintaining the pattern by calling this type an 'abugida,' from the Ethiopian word for the auxiliary order of consonants in the signary." However, some years later, the Abugida is another fixed Daniels' term passim in The World's Writing Systems, ed. by Daniels & Bright (1996).


acceptance in many publications, it also has its awkwardness. In his quest for the early history of the alphabet, such distinctions weaken the eye and obfuscate terminology instead of clarifying it. So the remaining question is: what must simply be recognized?

Granted that the distinction between Alphabet and Abgad were to be helpful from a typological and grammatological point of view, however, is it also correct or even useful regarding linguistics and Semitic philology? I am not so certain about this. Regarding Semitic writing, Daniels' distinction is artificial, and it is also dubious regarding Semitic philology and linguistics. Rather, this distinction is again to foster a certain bias as if the West Semitic alphabet were not yet entirely complete, lacking something important to be a fully working script system.22

In rejecting concepts of 'evolution' with respect to the history of writing, Daniels opts for "successive improvements" (Daniels 1996a: 8) . This though is awkward, too, because not every change in a system is necessarily

22 What is worse (however, Daniels is not to be charged with this) , the distinction in favour of a 'true' Greek 'alphabet' is capable of strange eurocentristic or at least graecocentristic chauvinistic effects, as Daniels himself pointed out by rebutting Eric Havelock (Daniels 1996b: 27-28) . Notions similar to Havelock's can also still be found in recent German publications, for example where of the Greek alphabet it is said that "die bahnbrechende Entwicklung eines Alphabets mit Konsonanten- und Vokalzeichen strahlte sowohl nach Osten aus [ . . . ] , als auch nach Westen [ . . . ] " (Casaretto, et al. 2007: 38), and Similarly: "Die Suche nach aLLgemeiner Kenntnis von den allgemeinen Gesetzen der Natur, die insbesondere der Medizin eine vollig neue und bis heute wirksame Fassung gab, war auf leistungsfahige Notationen angewiesen. Den Griechen kam zur Hilfe, dass sie ein gleichermai'jen exaktes wie flexibles Schriftsystem entwickeln konnten. Es forderte die Abstraktion, die man zur Erkenntnis gesetzesma�iger Vorgange in Natur und Gesellschaft benotigte" (Gerhard 2007: 6) . It goes without saying that this corresponds neither to basic historic facts nor to the historical deductions therefrom. Probably, the most radical German exponent of such a notion is the media theorist Friedrich Kittler (however, note the recent caveat by Jan Assmann 20n: no-m). Unfortunately, also the esteemed classicist Barry P. Powell (2009: 153ft) clings to a certain graecocentrism with fatal consequences for his perception of Semitics. Powell is at least poorly informed, when he contrasts "fewer than one hundred examples of West Semitic writing from the Levant" surviving from the whole first millennium BCE with the "epigraphic testimony to the Greek alphabet, swelling into an ocean of epigraphic remains" (2009: 153-154). Obviously, his volitional graecocentrism in matters of writing had let him astray regarding what the current state-of-the-art of West Semitic epigraphy really is (and already was during the past 50 years or more) . Actually Powell would had done better at least to ponder on what Tatian noted already in the second century: "Do not maintain a totally hostile attitude to foreigners, men of Greece, nor resent their beliefs. For which of your own practices did not have a foreign origin? The most famous of Telmessians invented divination through dreams, Carians foreknowledge through stars; Phrygians and the most ancient of the Isaurians the lore of bird-flights, Cyprians a cult of sacrifices; Babylonians astronomy, Persians magic, Egyptians geometry, Phoenicians education through the letters of the alphabet. Therefore stop calling imitations inventions." (Tatian, Gratio ad Graecos 1.1; Whittaker 1982: 2-3) .


an improvement of a former system, which then might be abandoned. Rather, at least in the most ancient Semitic alphabetical systems changes seem to be mere adaptations or adjustments to fit better the requirements of a given linguistic situation.

It remains problematic to take over the term Abjad in its 'Daniels' definition when dealing with the early history of the (West) Semitic alphabet. Even the late M. O'Connor got into confused terminology-coining terms like "southern linear abjad" (O'Connor 1996a: 98) or "South Semitic Abjad" (which nota bene denotes the Halal;ama) and wrote of "distinguishing the alphabet from the alphabetic order . . . to recognize the Levantine order and the South Semitic order, which must be independent of each other, as being of comparable antiquity" (O'Connor 1996b: 790) . Fatally enough, there was no unambiguous 'Levantine' order, as the rare but real Ugaritic Halal;ama evidence proves, but only an as yet statistically significant preference of Abjad (Abjad)-order use. Moreover, Daniels himself fails to discuss or even to mention the 'South Semitic' Halal;ama device that automatically would contradict him by its sole existence: one is curious about how terms would change if one day it happened that an Abjadsequence were found in Sabaic script within Sabaean boundaries, then perhaps speaking of hall;amic Abjads and abjadic (abgadic) Abjads . . . ?

To take a definition by David Crystal, from a linguistic point of view an alphabet simply is "a type of writing system in which a set of symbols (letters) represents the important sounds (phonemes) of a language" (Crystal 1992: 14 )-and one feels inclined to complement this with the second half of Daniel's definition of writing in general: . . . "in such a way that it can be recovered more or less exactly without the intervention of the utterer." (Daniels 2007: 55) . The sole point in question is whether a phonemic-graphemic alphabetic system works economically with its language ( s) or group of languages, which it was created for or adapted to. Or, to say it in other words: what are the important sounds (phonemes) , or what sounds were considered important by the majority of speakers or scribes of a language to recover an utterance more or less exactly without the intervention of the utterer at the time when a phonemic writing system for this language came into being?

The answer is clear: Greek is an Indo-European language with a large number of vowel phonemes and diphthongs and a lot of consonantal clusters, all of which are far more semantically distinctive and sensitive than in the Semitic languages. Thus, adopting the Abjad in the Aegean abets adapting it to the peculiar needs of their language type. By and by, assigning some graphemes of unknown sounds to vowels (some more

REINHARD G. LEHMANN

dropped out of use: koppa and digamma), and adding some more at the end, Greek finally created the Alpha-Beta, which fits almost perfectly as an

alphabetic base system for the non-Semitic Indo-European languages, and which finally made the game. This is not even rebutted by the few lndoEuropean languages that have been or are written with adaptations of Arabic or Hebrew script, because these adaptations include the additional and later vowel denoting tokens of these scripts as well.

On the other hand, by no means is the (west) Semitic Abjad a nonalphabetical system, and by no means is it deficient (Coulmas 2003: 113 ) . It is Daniels himself who concedes that "the Semitic abjads really do fit the structure of Hebrew, Aramaic, and Arabic very well [ . . . ] since the spelling ensures that each root looks the same through its plethora of inflections and derivations" (Daniels 1996b: 27 )-which is only partly true at all, if one regards such phenomena as assimilation and metathesis or Phoenician Sandhi at least in Northwest Semitic of the first millennium. The long Abjad system, enhanced by two or-depending on how one countsthree occasionally-used vowel signs (matres lectionis ) works perfectly until today with the Arabic language, as is proven by the (modern) Arabic script itself (and is indirectly and unintentionally proven even by Daniels 1997) , which has remarkably few differences from the Ugaritic long alphabet. And even the short Abjad, though with more frequent mater lectionis use, runs with Modern Hebrew. Thus, an Abjad truly is an alphabet in the linguistic meaning of the word.

Abjad or Alphabeta-it is mostly a matter of how much ambiguity one can tolerate and how much entropy one can stand before the writing becomes incomprehensible. "Given the systematic nature of [Semitic] consonant writing, it is clearly mistaken to look at it as something incomplete, an imperfection of technology which was to be fully developed only by the Greeks" (Coulmas 1996 : 92) . Recently, Jan Assmann also objected to connatural graecocentristic positions as those of the German media theorist Friedrich Kittler:

Wenn behauptet wird, die Vokalisierung eines konsonantischen Alphabets komme einer Kulturrevolution gleich, halte ich dies fUr eine Oberdramatisierung. (Assmann 20n: no)

As a set of letters, an alphabet is not the mirror of what should be there in a language from a phonemic or even phonological point of view, rather, it is the data stock of what provides maximum efficiency with least effort from a semantic point of view-which implies: the driver is the principle of least effort!


Anyway, the quest for the arrangement of letters is by no means irrelevant. Hence, the Abjad term, which as a numerological term originally denotes a certain letter sequence, remains an important terminus technicus exactly in this meaning, Le. , to denote the very letter order beginning with '-b-g-(IJ)-d-h-w . . . and should not be abandoned and ceded to grammatology. As a minimal terminological compromise, and to further avoid terminological confusion-which may be an effect of inconsiderately taking over Daniels' terms-I suggest using Abgad (with "g") to denote the Northwest Semitic alphabetic device ( see below) in contrast to Abjad, which unfortunately already seems to be established for a different notion in terms of grammatology.

Halal).ama or Abgad? How to Sort an Alphabet

As mentioned, the first basic parameter is the arrangement of letters, that is: the alphabetic sequence, Halal:zama or Abgad? In spite of some intrinsic variants and alterations, and in spite of certain exceptions, it is a matter of fact that, beginning with the first millennium BCE, one can ascribe both long abecedaries, the Halal:zama and the long Abgad device, to the two greater areas of their graphemic type, i .e . , the South Semitic and the North Semitic graphic letter sets respectively.23 Anyhow, it is noticeable and most interesting that the letter type that predates them both is the Ugaritic wedge alphabet. It is the Ugaritic cuneiform script alone that displays both devices, the Halal:zama and the Abgad, though by far not on an equal footing. Even though obviously the Halal:zama never was the standard in the realm of North Semitic and vanished in the Levant before the first millennium, and acknowledging the as yet unrevealed roots of the South Semitic script, it is almost indisputable that South Semitic Halal:zama and North Semitic Abgad are siblings (Trapper 1994: 298f.300, cp. Lemaire 2008: 50) . This becomes clear not only from their early coexistence in the Northern Levant in the second half of the second millennium BCE, but also from their common phoneme inventory.

23 However, note that there seems to be a certain 'zone of uncertainty' in the Ancient North Arabian, where letter inventories or abecedaries are found in (modified) Abgad, Halal:zama or simple order by graphic shape (Macdonald 1986, and infra n. 8) . This so far unparalleled diversity is explicable by the spread of literacy throughout " 'non-literate' nomadic communities which [ . . . ] thought of and used the ability to write in a quite different way to the sedentaries." (Macdonald 1986: llS and ff. )


For comparison, we arrange the South Arabic according to the (Ugaritic) long Abgad device:

Ugaritic

- '-b-g-tJ-d-h-w-z-/:l-t-y-k-s-l-m-cj-n-,?-s- '-p-s-q-r-t-g-t- 'i- 'u-s , -b-g-tJ-d-h-w-z-/:l-t -y-k-s-l-m-cj-n-,?-s- '-f -$-q-r-t -g-t -8-Q.

South Arabic / Sabaic (arranged in Abgad-mode)

It can easily be seen that there is a difference of only two or three signs respectively-the secondarily added Ugaritic signs in their normal position at the end, after t, and two more Arabic graphemes ( in the above chart these latter ones are not added at the end of the sequence, it seems so only because the 'Arabic' inventory is arranged here in 'Abgadmode' for better comparison) . The differences are indeed few, but they show that already in the late second millennium each must have had its own history.

Since the graphemic surplus of the Ugaritic 30-letter long Abgad is added at the end (letters 'i- 'u-s), and because there is a common phonemic inventory of 27 signs in both the Ugaritic and the South Arabic, although different in sequence, and because the South Arabic inventory has two more letters which are unknown in the Abgad tradition, it is most probable that they both represent an expanded superstrate alphabet, the basis of which must have been a stock inventory of 27, either as an Abgad, or as a Hala/:lama device.

Nevertheless, it remains unclear when the two devices separated from one another ( Schippmann 1998: 18-19)-if they where once united. The controversy becomes even harder because it applies to the basic question how the alphabet came into being at all and which of the two devices, the Hala/:lama or the Abgad, is the older one. But this is not the topic here.

Short or Long? How Many Letters Does an Alphabet Need?

The second parameter is the number of graphemes. Short or long? How many letters does an alphabet need? There is as yet no clear-cut archaeological and hard epigraphical data indicating that either the short alphabet or the twenty-seven-sign alphabet preceded the 22-letter standard alphabet of the first millennium BeE (Pardee 2003/04: 18 n. 45, Pardee 2007: 183f) .

Some scholars assume that the long Alphabet is an expanded version of the older short alphabet of 22 signs (enhancement theory, for instance


Dietrich & Loretz 1988: 141-143, 1991) . However, most scholars argue the other way round that the Ugaritic Long Alphabet (30/27) is an adaptation of an early form of the Northwest Semitic long Abgad device, which finally was reduced to 22 signs (reduction theory). The aporia is that so far there is no clear and unambiguous Northwest Semitic linear rich-inventory, i .e. , long (27/30) Abgad (or any other device) abecedary listed. Notwithstanding W.F. Albright's earlier claim (Albright 1950: 12-13) that there also must have been a 27-letter-linear long Alphabet, all attempts to unambiguously detect more than the well-known 22-letter linear Abgad have failed.24

But although the majority of the Ugaritic abecedaries is based upon the Abgad system, it is only Ugaritic that displays all types of Semitic alphabets as yet known to us: long (30-enhanced) Abjad, (27) long Hala/:lama, and short (22) Abgad. And it is notable that the full inventory alphabets (27/30) seem to be northern Levantine only, while the reduced-inventory texts are found at Ugarit, too, but more often to the south in Lebanon and Palestine, which means that it was scattered throughout the Levant.

Thus, to invoke Occam's razor, it seems most reasonable to adhere to the reduction theory. This posits that near the end of the second millennium in the Northern Levant an old, inventory-rich, alphabetic system of 27 consonant signs, which was almost exactly matching the consonantal inventory of the spoken language before, became reduced (Pardee 2007: 183f) .

As matters stand, i.e. , because there is not a single linearized instance of a long Abgad surplus letter, for instance It I or IIJI, we must also admit that by linearization of the graphemic type this reduction to 22 graphemes was already done, or in other words: the 22-letter short Abgad predates or is simultaneous with the emergence of linear letter forms, i .e. , the linearization. This implies that the reduction of the Abgad to 22 graphemes must have happened not later than in the third quarter of the second millennium, i.e. , most probably in the thirteenth century. This is confirmed by the date of the Ugaritic-written short Abgad texts.

As we now know, it was this trimmed down system that finally made the game. Its Levantine development and slight only alteration in the first millennium BeE gave birth to the Hebrew, the Aramaic, the Syriac, the Palmyrene, the Nabataean and others, and finally to the Arabic alphabet. By its transmission to the Aegean and further west it became subject

24 For instance M.F. Martin, 1962. A Twelfth Century Bronze Palimpsest, Rivista degli Studi Orientali 37: 175-193 (see infra note 18) .

REINHARD G. LEHMANN

to various alterations, modifications, and enhancements, but finally was spread throughout the world.

Shortening the Alphabet I: Who and When?

Admittedly, there is not a single primordial vowel sign in the Semitic alphabets. However, there is a need for grammatologists to understand that this is no impediment to understanding what an alphabet is. Also it is not the problem that we have a respectably high amount of letters, including vowel signs, in the Greek Alphabet. And it is also not a challenging question here why there are so many consonant -only letters in the South Arabic or in the Ugaritic writing tradition.

The most puzzling question is: What about the short Abgad system? Why are there so few letters left in

the Northwest Semitic Abgad of the first millennium BCE? Of course languages with a rich phoneme inventory can get along with a

deficient, or reduced alphabetic grapheme inventory (Coulmas 2003: 113) . On the other hand, "a borrowed alphabet in which phonemes of the borrowing languages are lacking tends to become polyphonic," (Blau 1982 : 3, cf. Blau 2010: 73ff)-which is only the reverse of the medal. For example, older Aramaic obviously had at least 26 consonantal phonemes (Creason 2008: 112-114) , but is written with the 22-Abgad system, and similarly this is the case with Hebrew for instance. Hence one cannot help admitting that the speakers of Hebrew or Aramaic adopted an incongruous alphabetical system, which beforehand had already been reduced to only 22 letters, insufficient to represent all consonantal phonemes.25 There again is no just cause to assume that their common donor or forebear never had more than 22 graphemes. The sole existence of a long Abgad tradition, though in cuneiform only, but nevertheless Abgad, makes a strong point against that.

But if a sudden reduction of the Abgad alphabetical system from 27 to 22 graphemes had taken place at some point more than 3000 years ago, where and when did this occur, and why was it done so? Most scholars assume that the (Proto-) Phoenicians did it, and there seems to be no reason to object to it (Lemaire 2008: 49f ) .

2 5 The same, to be sure, happened again more than a thousand years later when Arabic took over the Nabataean-Aramaic alphabet.


To be sure, there is no clear-cut evidence that the Phoenicians invented the alphabet, but there is less evidence that the Hebrews (or others) did. Thus, to invoke Occam's razor again, it seems most likely that the early Phoenicians were, if not the inventors of the 27-letter alphabet, at least the transmitters of the 22-letter alphabet. The very earliest texts of considerable length known to us that were written with the linear-short Abgad are undoubtedly Phoenician. These come from Byblos, like the famous Ahirom sarcophagus inscription (Lehmann 200sa, 2008b) or the Shipitbaal, Abibaal, and Elibaal inscriptions, and somewhat later also from other Phoenician sites like Tyre, Sidon, and the overseas colonies. Moreoverputting aside the inscribed Canaanite arrowheads of the late second millennium-, even the earliest of these texts already at the turn of the millennium display a considerable degree of scribal experience, skill, and craftsmanship (Rollston 2008a, Lehmann 2008b) . What we know of the original West Semitic texts, i .e. , inscriptions, from the middle or late second and even of the first millennium, are sherds, graffitti, and fragments incised in stone, but naturally nothing that is written flat. Flat writing however, i.e. , with ink on papyrus or another smoothed or planar surface, must have gone on simultaneously with and even earlier than clumsy scratches on sherds. This is also most likely considering the conjectural genesis of Ugaritic cuneiform (Lundin 1987b, Dietrich & Loretz 1988) . It is also evident that scratched or carved 'cursive' linear letter forms in argil always are predated by plain cursive types of professional flat writing (van der Kooij 1986) . Unfortunately, all this has gone forever because of the perishable nature of papyrus in the climate of Palestine, and, therefore, conclusions are speculative. The dearth of evidence for the early history of the linear alphabetic writing obscures our knowledge of the origin and early history of the alphabet. Nevertheless it has particular implications.

Some years ago, Benjamin Sass emphasized the sudden thirteenth-century alphabetic boom after a purported "excessively long palaeographic standstill" of ca. 1800-1300 BCE (Sass 2004-0S: 148) . This finally causes him to doubt the existence of (non-linear) alphabetic inscriptions preceding the fourteenth century, and to claim that the early history of the alphabet was not so early at all, but that the alphabet was born in Palestine in the thirteenth to twelfth centuries, shortly before it lost its 'Proto-Sinaitic' look and letter shapes became linear. However, notwithstanding the problematic dating of some early non-linear alphabetic inSCriptions, Sass' view is loaded with some methodologically problematic and romantic presuppositions, as, for instance, in this passage:


Rather than as a sudden 'boom', its timing quite inexplicable, ending a centuries-long freeze, this surge in the 13th-12th centuries may be understood as the swift, enthusiastic implementation of the breakthrough made in Egypt only shortly before. Brilliant inventions take root fast. (Sass 2004-05: 154)

However, by no means every 'brilliant invention' takes root as fast as Sass assumes, especially not if it undermines much older stable and established systems, as were the Mesopotamian cuneiform and the Egyptian writing. A 'freeze' or standstill in palaeographic development is only then problematic, if one confidently assumes that the bearers of an early alphabetic invention, and those who handed it down, were already 'democratic' in that sense that they shared their knowledge and skill freely with everyone who was interested. But how can we know whether they did so? The meagre contents of the few intelligible 'Proto-Sinaitic', or early nonlinear or proto-linear West Semitic inscriptions, including the recently published and highly controversial Qeiyafa ostracon (Misgav, Garfinkel & Ganor 2009), give no hint in this or in another direction. Hence, it is also possible that alphabetic writing was the arcane knowledge of only a restricted group of people or functionaries, such as some kind of (w )iiSipu( m) or early O�17 lawiyyim (ritual) specialists for instance, or other.26 As long as there was no broad social backing (and the evidence does not point in this direction), a new, phonematic 'bgd or hl/.zm based script system, which demands a good deal of abstract effort from its users, could not have 'boomed' suddenly.

Alphabetic writing is not such easy a skill to acquire in a short period of time-either by a gifted individual or even by a social group accustomed to writing, if any, in non-alphabetic systems, as often was assumed regarding the development of the Northwest Semitic alphabet. Rather, acquiring any fluent or even orthographic ally correct proficiency in writing a

first alphabetic system requires substantial time (Rollston 2008b: 68-69) . Moreover, special calligraphic knowledge, as it was already employed in the earliest Phoenician inSCriptions of Byblos (Lehmann 2008b) but is lacking in the 'Proto-Sinaitic' texts, needs a developmental period. This equally rules out an invention of the alphabet shortly before these inscriptions. Hence, one has to conceive a considerable incubation period in which the 'idea of alphabet' could steadily grow and stabilize as an effec-

2 6 Lemaire (2008: 49) speaks of "priests of local shrines who preserved and developed the use of this 'Canaanite' alphabetic script, perhaps so as to express a certain specific religiOUS and cultural tradition." From an ethnographic perspective, cf. also Schulz (1987) .


tive and competitive system, before it could achieve currency. "Thus, in the current state of the documentary evidence, the most credible working hypothesis would seem to link the origins of the alphabet . . . around the middle of the period 2000-1300" (Lemaire 2008: 47) .

No fast-taking root of some 'brilliant idea' was responsible for the boom and sudden spread of alphabetic literacy. Rather, we have to look for other 'booming factors' , which are equally able to explain an astounding-and maybe sudden-reduction of the alphabet in its Abgad device.

Shortening the Alphabet 11: Who Did What?

If it was the (Proto-) Phoenicians who boiled down the alphabet to 22 graphemes, and if their way of doing so immediately 'boomed' all over the Levant, why did they do so? We remember what the wise calligraphy master Serani said in Rafik Schami's novel: "The Arabic language has only twenty-nine letters, and the more of them you destroy, the more uncertain and imprecise the language becomes." Would it not be correct to say, that the fewer letters a script system has the more uncertain and imprecise the meaning? Semiticists know about the difficulties in deciphering Phoenician inscriptions because the defective and vowelless grapheme system doesn't even use supportive vowel letters, i .e. , matres lectionis.

The most prominent and traditional explanation seems to be that widespread linguistic changes had taken place in that period of time, which involved the loss of certain sounds. This is described as a typical feature of the Phoenician language. In grammars and concise descriptions of Phoenician, it is generally assumed that the ( spoken) Phoenician language had only 22 phonemes, which implies a full match of the graphemic with the consonantal phonemic inventory, i .e. , that Phoenician had as many graphemes as it had consonant phonemes. (Harris 1936: 16f; Swiggers 1991: 119f, Friedrich & Rollig & Amadasi 1999: 11; Krahmalkov 2001: 16.19,27 recently again Krebemik 2007: 124, Rainey 2007, MacDonald 2008b ) .28 The

27 However, Krahmalkov (2001: 19) notes that in early Phoenician of the Late Bronze Age the consonantal repertory was larger. Thus he self-contradictorily objects to the commonly held opinion that the 22-letter Phoinikeia grammata (ct'OlV lX� lc( ypalJ.lJ.c('TC( or XC(OIJ.� lc( ypalJ.lJ.c(Tc(, Herodot Hist. V 58-59) alphabet was invented by the Phoenicians but, rather, "had been devised in the Late Bronze Age and later" and was "adapted by the Phoenidans to the needs of their language which, in the early Iron Age, possessed a repertory of twenty-two consonantal phonemes" (Krahmalkov 2001: 16) .

2 8 Maybe one of the most pithy precis of this respect i s indeed made by MacDonald (2008b) , who states that "it was singularly unfortunate that the first widely used linear


assumption is drawn from the fact that the Phoenician alphabet-which is claimed to be the oldest standardized linear alphabet in the Eastern Mediterranean-has only 22 graphemes. And normally there are no further questions or doubts on that. Most notably, already in 1936 Zellig Harris, who was one of the most influential scholars in the twentieth century regarding Phoenician phonemics (on Harris, see Barsky 2011) , stated that "the complete absence of spelling variations in the Phoenician inscriptions as far back as they go would indicate that the alphabet conformed exactly to the needs of the language." However, Harris clearly saw and expressed the methodological weakness of the argument, admittedly also stating in an often overlooked footnote that "this is, of course, merely an argument from silence." (Harris 1936: 17 n. 21)29

The assumption equally implies that the Hebrews, the Aramaeans, and the other people of first millennium Levant speaking a non-Phoenician Northwest Semitic dialect, or language, unheSitatingly and invariably followed the Phoenician example in the trail of a supposed alphabetic 'boom' (Sass 2004-05) of the late second millennium BCE to content themselves with 22 graphemes only. But why should they have done so?

Alan Millard recently again pointed out that, "as ancient people read aloud, it may be assumed the signs gave sufficient information . . . " (Millard 2007: 85) . Texts were recovered through re-oralization, which means that the message was re-activated only by enunciating again that which was written down, thus, by reconversion of visible script into audible speech. Compare for instance the semantic development of the verb qr' 'to shout' > 'to read out loud' > 'to read'.

If this is true, the most intriguing point seems to be how the reconversion process from visible script to audible speech could have worked in an area, where a variety of presumably different Canaanite dialects with different phonemic treatments of the interdentals and the laryngeals still existed. It is widely accepted that the main Canaanite dialects including

alphabet was designed to express Phoenician, which had one of the smallest repertoires of consonantal phonemes of any Semitic language. Alas, the twenty-two letters of the Phoenician alphabet were treated as sacrosanct within the Near East, and the non-Phoenician Near Eastern languages which came to be written in it were squeezed into this rigid frame, regardless of the resulting ambiguities."

29 It should be also noted that already Harris made only one page later a possibly unconscious admission to a non-congruent phoneme-grapheme-relation in Phoenician writing: "The spelling in these inscriptions is etymological and the rules of orthography are rigid and unchanging, hiding all changes in pronunciation." (Harris 1936: 18, italics by R.G.L. ) .

Hebrew, including also the Transjordan fringe, and also the Aramaic varieties of the first millennium, despite of their 22-graphemes-only script system employed more than consonantal 22 phonemes in their speech at least up to the seventh or to the second century BCE respectively. The proof was already made by Bnmno 1970, Wevers 1970, Blau 1982, Degen 1969: 32ff, and Garr 1985, though with different methods, and many scholars had contributed to this topic in the last decades with further observations and adding more comparative material to subsequently refine the picture (Sivan & Cochavi-Rainey 1992; Hoch 1994, Steiner 2005) . Today, it is generally accepted that in the first millennium BCE for the vast amount of Northwest Semitic languages the 22-letter alphabet was deficient from a phonemiC point of view, or, in other words: the data reveal that at least to some extend sounds like the interdentals Iql, It I and It/ had not yet merged to their neighbouring sounds, and that often also the velar-pharyngal pairs of 1/J/-I/:lI and Igl-I 'I were still phonemically distinct and retained in speech at least in the late Bronze and Early Iron Ages-and further on.

It is true that Phoenician orthography remained conservative, rarely indicating vowels even after routine contact with the different devices to denote vowels in Greek, or, by use of matres lectionis, in Aramaic and Hebrew. But what, on the other hand, are the reasons to assume that the Phoenicians, only some hundred kilometres south from Ugarit with its rich phoneme and grapheme inventory, should have had a thus remarkably decreased consonantal phoneme inventory, dropping the interdentals and other sounds? Claiming that the Phoenician alphabet employed only 22 letters because that number represented the inventory of consonantal phonemes in their language means alleging that the 22-letter Abgad was a real phonemic (alphabet' , or system, at least for Phoenician. But is this really true? Since consonants carry the gist of information in Semitic languages, is it really feasible to say that the Abgad, once it lost some of its graphemes like </J>, «p, <q>, and <t>, was still purely phonemic? Or isn't that turning an outcome into its cause?

Granted also that there are no Phoenician texts in a 27-1etter alphabet, but only in a 22-letter system: is this sufficient reason to argue that they introduced the 22-letter short Abgad because they did not have more sounds in their language, whereas their eastern and southern Aramaic, Israelite, Judahite, and Moabite (recently Beyer 2010: lO-17) neighbours continued to articulate discernibly somewhat more than 22 phonemes in their dialect or language? Moreover, is it really within reason to imagine that those speakers simply jumped on the bandwagon of such mumbling

REINHARD G. LEHMANN

22-sounds-only-writing Phoenicians?30 From a methodological point of view there seems to be a considerable circular reasoning in these assumptions: it jumps too hastily to an explication, which anticipates the outcome (Phoenician script bequeaths only 22 graphemes) to be the cause (Phoenician speech had only 22 consonantal phonemes ) .

Even those who freely admit that there is inevitable material demonstrating that by no means was the early Phoenician language of the Levant a consonant-poor language and the 22-letter short Abgad simply displayed its actual phonemic state, do not try to escape the traditional view that the Phoenicians were responsible for the truncation of a formerly longer alphabet. Rather, one tries to postulate the short alphabet as the reflex of an actual "innovating speech of the royal elite", thus introducing a 'short speech' as a kind of sociolect that finally "would also be used for the speech of the common citizens" (Hoch 1994: 415) . However, this is a way to subsequently reduce the derivation group of the short Abgad by and by into nothing: which royal elite where?

Anyhow, a last retreat from which to argue the point of a 22-letter oneto-one grapheme-phoneme correlation again and again still seems to be Phoenician, which is held responsible for the spread of the short Abgad even over an area where languages with more than only 22 phonemes still were spoken. Nevertheless, these languages continued for quite some time, possibly for the duration of their existence, to write and spell in accordance with a presumed phoneme-based Phoenician-hence for themselves deficient-orthography.

Shortened Phoneme Inventory: How Many Phonemes?

If it is true that Phoenician and the Phoenicians were responSible for the spread of the Short Abgad system (and the epigraphic evidence from Byblos for instance still points in that direction), there is another difficulty that has come to the attention of scholars in the last decades: is it

30 Accordingly, Rainey (2007: 69) from an entirely different approach and with a view to 'redefine' Hebrew as a Transjordanian language simply claiming that "the speakers of Hebrew did not speak the same dialect as those from whom they borrowed the alphabet", recently states that these borrowed the Canaanite ( = Phoenician) alphabet because "the rustic clans from the steppe lands were so impressed by that superior cultural feature that [ . . . ] they adopted the writing medium of the highly cultured people of the coastal areas." Rainey correctly points out the question of an obviously deficient graphemic system for these languages. However, his explanation seems to be as romanticized as that of Sass.


still possible to argue that Phoenician had 22 phonemes solely because the 'Phoenician' alphabet had only 22 graphemes? This became one of the key assumptions for the reconstruction of Canaanite. However, the underlying syllogism is not compelling, and by no means is it mandatory or self-evident that early Phoenician should have had 22 consonant pho- . nemes only.

Without any claim to be comprehensive, let us have a look on some of the relevant data, mainly from Egyptian sources. The transcriptions of Semitic words in Egyptian writing of the New Kingdom and the Third Intermediate Period as discussed by Hoch 1994

probably tell us more about old Phoenician and Hebrew than current hypothetical reconstructions of these languages can tell us about the origin of the words here studied. The generally accepted reconstructions rest on a

very bold assumption-that the short Semitic alphabet contained a grapheme corresponding to each phoneme. It will be suggested that the assumption that the bulk of Semitic words are from older dialects of Phoenician and Hebrew is more likely than the assumption upon which Phoenician is reconstructed. The conclusion to be drawn is that our current reconstruction should be reconsidered. (Hoch 1994: 12)31

From Hoch's data it follows that "the most reasonable conclusion seems to be that most of the Canaanite dialects had a fairly extensive phonemic inventory and used a short alphabet, with polyvalence of several graphemes". His data confirm the polyvalence of certain short Abgad letters in the Late Bronze and Early Iron Ages as employed by Virtually all Canaanite dialects, and even so that "Phoenician may have also been more conservative in its phonemic inventory than universally believed" (Hoch 1994:

416-417 ) .32 However, because the source language of Semitic words in Egyptian

writing as discussed by Hoch 1994 remains unsure in many cases, only the few well-defined unequivocal Phoenician examples are listed here.33 Thus, from a methodological point of view, two conditions are to be applied

3 1 Compare also, towards the end of Hoch's book: "By far the most important factor that emerges from the Egyptian evidence is the size of the phonemic inventory, which numbers as high as 27-29 phonemes, even in the more recent material. This is far more than usually believed to be present in the contemporary Canaanite dialects. Although it cannot be demonstrated that any single dialect contained the full inventory, it would seem likely that dialects with mergers were in the minority as source languages" (Hoch 1994: 413) .

3 2 Cp. also Krahmalkov (2001: 19), however, see infra note 27. 33 More examples, though not unequivocally classifiable as proper Phoenician, are

found in Helck (1971: 507-527); Sivan & Cochavi-Rainey (1992); Schneider (1992) ; and Hoch (1994) passim.

REINHARD G. LEHMANN

here to Hoch's material, which causes a substantial restraint of evidence: a) the Semitic foreign word must comprise diagnostic consonants to prove a polyvalence of the Short Abgad at all, and b) the Semitic foreign word must be identifiable as unequivocally Phoenician by external evidence, thus being a Phoenician toponym or anthroponym, or by the shift of short accented *Ial > 101 .34 Anyhow, even if every word which is not well-defined Phoenician by these conditions is ruled out, there are still enough clear-cut arguments left over:

In the Egyptian story of Wenamun (1,29; 3,7 and 1,16-17; Schipper 2005: 177) , the name of the Phoenician-Byblian ruler Zakarba cl, or Zakurba cl, is written T-k-r-b- ' -l representing the voiced interdental phoneme Idl, clearly distinguishing it from Izl (Schneider 1992: 256f; Hoch 1994: 372-373, 400 n. 5, 417) . This retention of Idl in the twelfth or eleventh century gives witness that the Phoenicians still retained a phonemic difference between Idl and Izl at a time when the short Abgad already was in use.35 Nevertheless, on at least three almost contemporary or slightly later Phoenician arrowheads36 the orthography of this name has <z> , which was the only possible orthography in a Canaanite-based 22-letter Abgad system, as is also displayed in somewhat later cognate Hebrew names like i1;l�T Zakaryii, or �i1;l�T Zakaryiihu, and the Aramaic hypocoristicon '�:l! Zakkiir. Because the root *dkr ( 'to remember' ) not only in Phoenician, but also in some earlier Aramaic texts is written zkr ( in contrast to later Aramaic dkr), it is most probable that at least until the early first millennium even in Phoenician Canaanite the overall phonemic polyvalence of the short Abgad <z>-letter was retained.37

Also the original interdental * It I seems to have been still phonemically different from 1$1, as Egyptian writings of the roots 'n; (* 'rt) 'to terrify', r$ ( <*rt, Hebrew 1" ', 'to go swiftly) , and the noun $m' « *tm', Hebrew NO�,

'thirst' ) suggest (Hoch 1994: 405) , though we cannot be sure whether the source language of these words was Phoenician. However, again there is at least one clear example: The name of the Phoenician city Tyre (Hebrew

34 The Canaanite shift, however, is to handle with care because it could be diagnostic to Canaanite in general, but seems to be ruled out here anyway because of the vowelless Egyptian script.

35 Anyhow it should be mentioned that some object to such an optimistic differentiation between the Egyptian writing of Semitic <rj> and <Z>, Quack: ZDMG 146 (1996) 507-514.

36 See Starcky (1982: 180) and Deutsch & Heltzer (1995: 24-26). 3 7 Schneider (1992: 257) additionally mentions the name t-k-r-m-w of a person stemming

from Gaza which, of course, cannot be collected as Phoenician proper.

1� $or) is given in Wenamun l)-r (1,28; 3,6; Schipper 2005: 178£), still phonemically representing its original and etymological sound It I as in tr.38

Further, owing to the Phoenician literary ambience of its records in the story of Wen am un, it is obvious that also the phonemic IIJI-I/;I opposition

· was maintained in Phoenician. There are at least two examples:

ma=sa=IJi (Wenamun 1,9-10; Hoch 1994: 152, Schipper 2005: 45-46) for Phoenician *maSiIJa 'amphora', compare Hebrew ;'1)o/Q 'anointing', and IJ=b� =-r (Wenamun 2,1 and 1,X+24; Sivan & Cochavi-Rainey 1992: 12f; Hoch 1994: 240; Schipper 2005: 65.187.189£) for Phoenician *IJabira 'business, trading partner, colleague', compare Hebrew 1;tt:), 1;1.1) 'companion' and 1-i-TJ 'guildsman'.

Thus, from this sparse evidence alone there is sufficient reason to conclude that in the first millennium BCE Phoenician, like Hebrew, had more than 22 different consonants and at least two more semantically sensible phonemes ( l1J1 and lu/) , if not more, even in later times: A close analysis of the Persian period Yahawmilk inscription from Byblos, for instance, shows that rules of a written 'Space Syntax', which is a background radiation of oral performance, consistently and on three occasions served to 'save' the minimal pair ptIJ "to engrave" and pt/; "to open", avoiding a consonant sandhi where I IJI and I/;I clashed on word boundaries in the construct-genitive relation ptIJ /;r$ "golden engraving" (not: "golden door", Lehmann 2oo5b: 88-90) .39 The semantically distinct phonemic values IIJI and I/;I for the one only common grapheme </;> were still in use. And maybe there were more. Moreover, it is not very plausible to assume that only 200 km south of Ugarit and either coeval with or only 200 years after the long Abgad, in Phoenician a sound system simply and almost

38 Because of the "key fact" that the Egyptians generally perceived a difference between the Canaanite reflexes of original * W and * /s/ (Hoch 1994: 404), in principle also this could be a source to prove that Phoenician had more phonemes than only the 22 that are presented by the short Abjad. Unfortunately, as far as I can see there is no well-defined, unequivocal Phoenician example for that. However, there is the often refuted (cp. Garr 1985: 28) note of Piu tar ch that the Phoenician word for "bull" was 6wp instead of expected (J'wp, which could reflect the still interdental pronunciation tor: LWP yap Ol <PO[V lXE� TIjv �ouv XctAOU(J'l (Vita SuLLae xvii) .

3 9 The known Phoenician notion to produce graphemic sandhi-such as in mlqrt, the deity "Milqart", from mlk-qrt "King-of-the-city" (Friedrich-Rollig-Amadasi 1999: 48.56), and more (Schmitz 2011)-presumably could be one scholarly reason to insist in assuming that already in the late second millennium Phoenician had an intelligibly reduced phonemic inventory of 22 consonantal sounds, thus also to afford 22 graphemic units only in the linear alphabet. However, also vice versa it is conceivable that a principle of scriptio franca (vide infra), once being established, abets to increased productivity of spoken and written sandhis.

immediately shrinked down to only 22 consonants, whereas Ancient Hebrew, the next contiguous language,40 retained the phonemic opposition of Ibl versus l/:ll and Igl versus / , 1 until well into the first millennium, and probably up to the second century BCE, as can be proven by the Septuagint and other transliterations for instance (Wevers 1970, Blau 2010: 75-76.86, Steiner 2005) .41 Similarly the Aramaic languages of the first millennium retained older, not any longer graphemically represented phonemes such as voiced interdental I dJ, and also an uncertainty in the representation of the emphatic voiced interdental * If/I by either older <q> or younger < '> until well into the first millennium (Folmer 1995, passim) .

Shortening the Alphabet Ill: Why? And What Was It Good For?

How than is it any longer possible to explain the shortening of a long, almost phonemically congruent script system (the long Abgad) into a deficient, polyvalent script system (the short Abgad)?

To begin with, I would like to emphasize three main aspects again regarding the short Abgad:

- The Ugaritic short alphabet is attested mainly late (thirteenth century BCE) and not only at Ugarit, but also scattered around the Levant and Cyprus. Its writing direction is not (yet) fixed.

- The Greek alphabet has consistently not a Single feature that could be accounted to the long Alphabet tradition, neither to Ugaritic cuneiform nor to any Canaanite long-linear alphabet, and neither to the Hala/:lama series nor to an assumed linear long-Abgad device.

- The Greek alphabet, however, is obviously a dependant, if not offspring, of the 'Phoenician-style' linear short Abgad series. This is clear both by the way of its various 'inner' modifications and by its enhancement only at the end, i .e. , after < t> (Wachter 1996) .

40 To be sure, here the term 'Hebrew language' i s used for convenience only, bearing in mind a South Canaanite dialect cluster of the late second and early first millennium BCE which eventually evolved to what only later and tentatively could be labelled as 'Hebrew' language (Ullendorff 1977, Knauf 1990, Lehmann 2008a: 6-16) .

41 Similarly even Steiner (ZOOS: 261), who opts for Phoenician influence being responsible of the later Hebrew merger of Ib./ and Ib.!, wonders about "the glacial pace of the diffusion, with the merger taking more than a millennium to reach Jerusalem" If, however, the conjecture of this paper is correct that there was not an early phonemic merger of these sounds in Phoenician at all, there is nothing to be worried about.


Thus, the 'Phoenician-style' short Abgad alphabet seems to be most probably developed in the course of the partial 'globalisation' of the Eastern Mediterranean by the Sea Trade in the late second millennium, Le., in the thirteenth century BCE (Artzy 1997, 1998, Brody 1998, Haider 1988-1989, Negbi 1992, Sherratt 1993, 1998, Schipper 2005: 113-138, and Monroe 2009 ) . This becomes even clearer if considered the trade connections as described in the (to be sure, later) trade list of Ez 27, esp. 12-24 (Saur 2008: 107-225) . The 'Phoenician'-type short Abgad system became what one may label a scriptio franca of the Eastern Mediterranean.

This was a process similar to texts written in the Phoenician language being a regional standard lingua franca even in regions that presumably have not been colonized by Phoenicians, such as Karatepe in Anatolia, Hassan-Beyli, Cinekoy and Sam'al (Greenfield 2001: 179, Krebernik 2007: 118, Lemaire 2008: 52 ) . Note that my use of scriptio franca refers only to the Phoenician alphabetic writing system, not necessarily to the Phoenician language or any other dialect in the Canaanite group. Nevertheless, the Phoenician-Levantine dialect continuum gave birth to, and provided a scriptio franca.

The shorter and less sophisticated a script sign inventory is, the easier it adapts to other language systems or dialects respectively. In assigning phonemic values to their graphemic token we must be aware of regional differences, and even very close and spatially limited regional differences. Not every single difference in pronunciation is clearly discernible, but undoubtedly these variations were there. The classic and famous Biblical "Shibboleth" incident of Judges 12:16, where the Ephraimites did not conform to the expected pronunciation of the word sibb6l�t, rather saying sibb6l�t or the like (which is not necessary to discuss here in detail) ,42 is only the most prominent tip of the iceberg.

Moreover, there is the 'summer fruit/end' pun in Amos 8:2, which plays with the Hebrew words qayi$ 'summer (grain/fruit) , (written 'P"P qayi$) and qe$ 'end' (written 'PP qe$) :

42 See, for instance, J. Marquart, 1888. Sibbolet = ephraimitisch Sibbolet? ZA W 8: lSllSS; P. Swiggers, 1981. The Word Sibbolet in Jud XII.6, jSS 26: 20S-207; E.A. Speiser, 1942. The Shibbolet Incident (Judges 12,6) , BASOR 8S: 10-13; Gary A. Rendsburg, 1988. More on Hebrew sibbolet,jSS 33: 2SS-2S8; Andre Lemaire, 1985. L'incident du sibbolet (Jg 12.6) . Perspective historique. In Melanges M. Defear, (AOAT 21S) , pp. 27S-281, Neukirchen; Ronald Hendel, 1996. Sibilants and sibbolet (Judges 12:6) . BASOR 301: 69-7S; J.A. Emerton, 1985. Some Comments on the Shibbolet Incident (Judges Xn.6) . In Melanges M. Delear, (AOAT 21S) , pp. lS0-lS7, Neukirchen; and others. For the sibilant distinction and polyphony of the Hebrew iV-sign generally, see, most recently, Blau (2010: 73-7S).

42

ma- 'atta rij '� 'amos wa' omar k3Lub-qayi� wa'y'yom�r YHWH 'elay

REINHARD G. LEHMANN

ba haq-qe$ '�l- 'ammi yisra' eL

Amos, what seest thou? And I said: a basket of summer fruit. Then the Lord said unto me, The end is come to my people Israel

In the northern Hebrew dialect full monophthongization seems to have taken place, which blurs the difference between f"i' qayi$ and fi' qe$ anyway, thus leading to fi' qe$ in both words.43 Insofar Blau 2010: 8 is right when he states that "the pun would have been more powerful in the speech of the population of the Northern Kingdom, whom Amos was addressing, if Amos pronounced qayi$ as qe$." However, this could not have been a sufficient reason alone to make the pun work, because in this case his addressees would have understood the last line as summer has come to my people Israel-which is not thus unexpected at all and normally happens every year. Therefore there must have been a further and more stable pronunciation difference between fi' ( 'summer' ) and fi' ( 'end') beyond writing alone to make the pun play properly any effect. Certainly this was the difference between the emphatic voiceless fricative dental in fi' /qe$/ 'end' and an assumed emphatic voiced fricative dental in fi' or f"i' /qe$/ 'summer', as exemplified in written Ugaritic q$ 'summer' for instance.

Also, a third demonstrative example, though Aramaic, should be given from the Old Testament. It is the single dispersed Aramaic verse of Jeremiah 10:11:

'IELahayya di s3ma,Y.Ya w3- 'arqa la 'abadu yebadu me'ar'a umin t3/:lot s3ma,Y.Ya 'ell�

The gods that have not made the heavens and the earth, these shall perish from the earth and from under the heavens

Besides the phenomenon of a chiastic word play, it is conspicuous that the same word for 'earth' is written differently in the first and in the

43 For instance line 7 of the tenth-century BeE Gezer calendar inscription, where the "month of summerfruit" is written yrlJ-q$, or the word "wine" in the (northern) 'Israelite' Samaria Ostraka and R. ez-Zetun, where it is written l' /yen/ in contrast to instances in (southern) 'Judahite' inscriptions (Arad, Lachish, el Qom), which are constantly written r� /yayn/ (even in the construct state! ) in accordance with Biblical Hebrew orthography.

second stichos, i.e., within an interval of few words only: the first instance is written 'arqa (in consonants only: 'rq'), while the second is written 'ar'a Cr") . Similar examples are found in some Achaemenid imperial Aramaic papyri of the same period, i .e. , the fifth/sixth century BCE.44 This orthographic interchange of the same word in one single verse could only be tolerated because in that very time period both writings were perceivable as the same word, i.e., 'earth', which in older Aramaic most regularly is written with the letter <q> , and in later Aramaic with the letter < '> . This was possible because of the etymology of the word, which is old /' an/i, the phoneme Idi which the Levantine Abgad dropped when it became reduced to 22 graphemes (a similar example is the inner-Hebrew dialect variant m/:zq "to smite" adjacent to standard Hebrew-Canaanite m/:z$ in the same verse Judges 5:26) . However, still in the middle of the first millennium BCE there was such broad a range of pronunciation for this sound possible that in a transitional stage of one or two centuries it could be rendered by the grapheme <q> or < '> of the 22-letter Abgad respectively without causing serious problems for the meaning of the word and even the understanding of an utterance. Despite the different orthography, on the other hand, the pronunciation could not have been all too seriously different, rather at least similar.

The Canaanite Semitic languages or dialects of the late second millennium Levant were by no means incompatible. Rather they form a dialect continuum (which excludes Aramaic, to be sure) that reached from the Northern Syrian coast down to the South Palestinian seashores (maybe excluding some Philistine areas, admittedly), and from Galilee throughout the Jordan valley up to the Golan and Basan heights in the east (Garr 1985) .45 Considering this, and given the need of a script standard for supra-regional trade affairs, one should try to imagine what would have happened if every shift of articulation in regional speech were to be represented in a written document using a 27-or-2g-consonant system such as the Ugaritic long Abgad or the Sabaic Hala/:zama device? The Levantine Northwest Semitic languages display such an enormous number of lexical isoglosses, which all too often differ from a graphemic point of view only

44 For instance in Aramaic documents from Egypt, TAE B2.2:5.12.15 (464 BCE) 'rq', in the same document only one line later (B2.2:16) 'r". In the eastern periphery the older spelling with <q> did survive until the third century BCE in the Aramaic Asoka inscriptions from Kandahar, and others (Folmer 1995: 63-70) .

4 5 Not to imply that this book is exempt from methodological criticism.


(Halayka 2008) , that the major requirement was not for a supra-regional standard language.

Rather, it deserved a graphemic system that was able to blur the articulation and pronunciation differences for the purpose of East-Mediterranean and Levantine mercantile affairs, a scriptio franca.

Once again to recall Alan Millard (2007: 85) : "as ancient people read aloud, it may be assumed the signs gave sufficient information . . . ". Because texts were recovered through oralization, i .e. , by reconversion of visible script into audible speech, at first glance it seems to be plaUSible that aside from any administrative centre or official use at a royal court or scribal school using a high literate standard, alphabetic writing in the Levant can only mean this: The representation of the oral language by means of graphic characters, i .e. , the graphic representation of any oral autochthonous language. But whoever coined an alphabet, or shrank an inherited one, was necessarily limited by his own perception of sounds or, to say more, he was restricted by his way to interpret slightly different sounds as representing the same phonemic notion in principle, which means to hear, or to perceive, or to interpret the sound /1/ as even in some cases representing also /s/ (or /t/, vice versa) , the sound /d/ as /z/, the sound /v/ as //:t/, and so on, melting different areal dialectal pronunciation varieties into one.

However, on the other hand, what would have happened if every sailor or tradesman in the eastern Mediterranean sea trade made his records by tracking down his own regional Canaanite-Semitic tongue, or articulation, in that rich consonantal graphemic system (the Long Abgad), which graphically differentiates between sound like /t/, /1/ , and /s/, or /v/ and //:t/, or between / '/ and /[j/, between /d/, /d/ , and /z/, and so on? The basic question was risen already for instance by Childe 1942 : 181.182:

. . . the merchant would have to be his own bookkeeper. This was the social background of the Phoenician script. [ . . . ] Twenty-two signs were chosen to denote simple consonants-vowels were not written. [ . . . ] It was in fact an international body of merchants who sanctioned by use the new conventions; it was their activity that diffused and popularized the system in the Iron Age.

Notwithstanding that Childe did not yet reflect on the question of the difference between long and short alphabetic traditions, but rather adhered to the position that shortening the Abgad has had its reason in a reduced phonemic inventory of the Phoenician language, he made a good pOint: i .e. , to emphasize "the mercantile applications of writing that were needed


to excel in long-distance trade" (Monroe 2009: 144) .46 To conclude in one sentence: a standard script system was required for use in supra-regional or quasi-international, i.e. , eastern Mediterranean trade affairs, but, to be sure, not that much a regional standard languageY

And what about Aramaic, which does not belong to the Canaanite dialect continuum? Aramaic also took over the 22-letter Abgad system, but can by no means simply be subsumed under that group of languages that formed the economic linguistic standard of the Eastern Mediterranean sea trade-neither by linguistic reasons, nor from a geographical point of view. Hence it is notable that linear alphabetic written Aramaic, at its earliest attested level, ran remarkably close together with Phoenician even there where Phoenician or any contiguous Canaanite dialect was not the vernacular. In a similar way as Phoenician seems to have been a

second-level regional standard next to Luwian in Anatolian Karatepe and the kingdom of Que (Lemaire 2008: 52£), Phoenician was also next to Aramaic in Zincirli (ancient Sam'al) , where a supposedly Luwian ruler named Kulamu mainly used Phoenician for his official declamations. But already his successors took over their vernacular Aramaic language into the Phoenician script system, subsequently forming their own Aramaic script standard on the basis of the 22-letter-Abgad. Thus, the starting point of the use of the 22-Abgad system by Aramaic local rulers was by clear Phoenician influence, maybe to engarland themselves with an international looking, global-player-like flair.

Conclusion

As a scriptio franca, the 'Phoenician' -style 22-Abgad of the late second millennium became the historical bottleneck of alphabetic development. This happened most probably by means of mercantile requirements of the Levantine sea trade, which dominated international affairs of the Eastern Mediterranean for a while, at least as far as the global players of traffic and trade were concerned. Some hundred years later coming out of this

46 To be sure, Monroe does not refer to Childe here explicitly. He rather draws a comparison to mercantile use of literacy in mediaeval Italy.

47 Maybe within here lies the main difference to Young (1993: 4-26) who basically focuses on Phoenician as a prestige language. However, see Young (1993: 25) : "The possibility must be borne in mind that our written sources may reveal much more of local scribal practice than of genuine grass-roots differences in the spoken language."

REINHARD G. LEHMANN

bottleneck again, the Phoenician short Abgad already had made the game and would remain an international prestige script of the Levant for some time (Naveh 1987; Young 1993: 4-26; Rollston 2008b: 72, 78, 89), before the alphabet( s) evolved into different 'national' script systems. These could not track back any longer to the former functionally rich 27-phonemeinventory, though more and different phonemes might have been needed to properly display the vernacular. Rather, because of that true 'bottleneck effect', those 'national' scripts had gradually to evolve their own way to graphemically display the effects of interdental pronunciation of * Idl either by <z> in Hebrew or <d> in Aramaic for instance,48 and also to introduce their own tokens for vowels, vowel-length, and supra-segmental divisions in Ancient and Classical Hebrew, Aramaic, and other Semitic languages, as well as in Greek. For these different 'national' scripts then, which generally can be distinguished as the scripts of politically and / or culturally defined population groups, it was not any longer in the trade where the script was spread, but through the courts or political centres.49

This hypothesis hopes to explain: The starting point of the global success story of the 22-letter-Abjad. The multilingual power or ability of the 22-letter-Abgad (Northwest Semitic and beyond, and early Greek) The performance of the 22-letter-Abgad even in North Semitic languages that eVidently had a richer phoneme inventory than the graphemes of the 22-letter-Abgad (Aramaic, Hebrew, and also Phoenician) The launch of matres lectionis only in the realm of the 22-letter-Abgad, which re-enhance a partially deficient consonantal phonemic system, in contrast to the South Arabic Hala/:zama series (MacDonald 2008b), which retains a more rich consonantal inventory of 29 phonemes to endure lexemic differences.

48 More examples are given in Rollston (2008b: 65) for instance, and in many introductions into Northwest Semitic languages. For Aramaic, note that Old NW Semitic * It I was represented by <s> in Old Aramaic, sometimes by <S>, and later in Imperial Aramaic by <t>, proving also that the phonemic variation of the interdentals was active in speech at least in the first half of the first millennium BCE.

4 9 This sentence picks up van der Kooij (1986: 250-251), but with important modifications. The original sentence quotes : "Moreover, the regions can generally be distinguished as politically and often culturally defined population groups. This points to a politically centralised teaching of writing at one or more interrelated royal courts." (p. 250) , and: "To sum up, the tentative conclusion is that the wide nib ink writing was the one taught almost exclusively, and that it was not in the trade, nor in the army, nor in the religious institutions that the script was spread, but through the courts or political centres." (p. 251) .


The survival of the Hala/:tama series in South Semitic only, which was not in the sphere of influence of the Levantine Sea Trade. The early independent adaptation and subsequent enhancement of the 22-letter-Abgad in the Aegean and further west by mercantile requirements.

References

Albright, W.F. 1950. Some Important Recent Discoveries: Alphabetic Origins and the Idrimi Statue. New Light on the Beginnings of the Alphabet. BASOR u8: 12-14.

Artzy, Michal. 1997. Nomads of the Sea. In Stuart Swiny, Robert L. Hohlfelder, Helena Wylde Swiny (eds.) , Res Maritimae. Cyprus and the Eastern Mediterranean from Prehistory to Late Antiquity. Proceedings of the Second International Symposium "Cities on the Sea)� Nicosia, Cyprus, October 18-22, 7994, pp. 1-16. Atlanta, Georgia: Scholars Press.

--. 1998. Routes, Trade, Boats and "Nomads of the Sea". In S. Gitin, A Mazar, E. Stern (eds.) , Mediterranean Peoples in Transition. Thirteenth to Early Tenth Centuries BCE, pp. 439-448. Jersualem, Israel Exploration Society.

Assmann, Jan [& Frank M. Raddatz) . 20ll. Pharaonendammerung. Irrungen der Unsterblichkeit - vom tragischen Heros zum Gotteskrieger. Lettre International 92: 104-ll4.

Barsky, Robert F. 20ll. Zellig Harris. From American Linguistics to Socialist Zionism. The MIT Press.

Bernal, Martin. 1987. On the Transmission of the Alphabet to the Aegean Before 1400 B.c. BASOR 267: 1-19.

--. 1990. Cadmean Letters. The Transmission of the Alphabet to the Aegean and Further West before 7400 B. C. Winona Lake: Eisenbrauns.

Beyer, Klaus. 2010. Die Sprache der moabitischen Inschriften. KUSATU ll: 5-41. Blau, Joshua. 1982. On Polyphony in Biblical Hebrew. Proceedings VI, no. 2. Jerusalem: The

Israel Academy of Sciences and Humanities. --. 2010. Phonology and Morphology of Biblical Hebrew. An Introduction. Linguistic Stud

ies in Ancient West Semitic, 2. Winona Lake: Eisenbrauns. Bordreuil, Pierre. 1982. Quatre documents en cuneiformes alphabetiques malconnus ou

inedits: Semitica 32: 5-14. Bordreuil, Pierre & Dennis Pardee. 1995. Un abecedaire du type sud-semitique decouvert

en 1988 dans les fouilles archeologiques franc;aises de Ras Shamra-Ougarit: Comptes Rendus de L'Academie des Inscriptions et Belles-Lettres, 1995: 855-860.

--. 2001. Textes alphabetiques en ougaritique. 8. Abecedaire: Etudes Ougaritiques 1. Travaux 7985-7995 (RSO XIV), 2001: 341-348.

Brody, Aaron J ed. 1998. Each Man Cried out to his God. The Specialized Religion of Canaanite and Phoenician Seafarers (Harvard Semitic Monographs 58 ) . Atlanta, Georgia: Scholars press 1998.

Bron, F. & Ch. Robin. 1974. Nouvelles donnees sur l'ordre des lettres de l'alphabet sudsemitique, in: Semitica 24, 1974, 77-82.

Bnmno, Einar. 1970. Die Aussprache der hebriiischen Laryngale nach Zeugnissen des Hieronymus, Aarhus 1970.

Byrne, Ryan. 2007. The refuge of Scribalism in Iron I Palestine: BASOR 345, 1-31. Casaretto, A & A Daues & J.L. Garda-Ram6n & AV. Sansalvador. 2007. Vom Bildzeichen

zum Buchstaben. Schriftzeichen im antiken Mittelmeerraum des 2. und 1. Jts. v. Chr., in: Dietrich Boschung and Hansgerd Hellenkemper (eds. ) , Kosmos der Zeichen. Schriftbild unfBildformel in Antike undMittelalter (Schriften des Lehr- und Forschungszentrums fur die antiken Kulturen des Mittelmeerraumes - Centre for Mediterranean Cultures, Band 5 ) , Wiesbaden 2007, 21-43·

REIN HARD G. LEHMANN

Childe, V. Gordon. 1942. What Happened in History. Harmondsworth: Penguin 1942. Coulmas, Florian. 1996. The Blackwell Encyclopedia of Writing Systems. MaIden, Oxford:

Blackwell Publishing 1996, 1999. --. 1989. The Writing Systems of the World. Oxford: Basil Blackwell 1989 (reprinted

1990.1991). --. 2003. Writing Systems. An Introduction to their Linguistic Analysis. Cambridge Text

books in Linguistics. Cambridge: Cambridge University Press 2003. Creason, Stuart. 2008. "Aramaic", in: Roger D. Woodward (ed.) , The Ancient Languages of

Syria-Palestine and Arabia. Cambridge University Press 2008: 108-144. Cross, Frank Moore. 1967. The Origin and Early Evolution of the Alphabet: El 8: 8*-24*. --. 1980. Newly Found Inscriptions in Old Canaanite and Early Phoenician Scripts:

BASOR 238: 1-20. Crystal, David. 1992. An Encyclopedic Dictionary of Language and Languages. Oxford. Daniels, Peter T. 1990. Fundamentals of Grammatology.JAOS 110: 727-731. --. 1996a. Grammatology. The Study of Writing Systems. In Peter T. Daniels & William

Bright (eds.), The World's Writing Systems, pp. 1-17. Oxford: Oxford University Press. --. 1996b. Ancient near Eastern Writing Systems. The First Civilizations. In Peter T.

Daniels & William Bright (eds. ) , The World's Writing Systems, pp. 19-32. Oxford: Oxford University Press.

--. 2007. Littera ex Oriente. Toward a Functional History of Writing. In Cynthia L. Miller (ed. ) , Studies in Semitic and Afroasiatic Linguistics Presented to Gene B. Gragg, pp. 53-68. Chicago.

Degen, Rainer. 1969. Altaramiiische Grammatik der Inschriften des 19.-8.Jh. v. Chr. Abhandlungen fur die Kunde des Morgenlandes 38(3) . Wiesbaden.

Deutsch, Robert & Michael Heltzer. 1995. New Epigraphic Evidence from the Biblical Period. Tel Aviv-Jaffa.

Dietrich, Manfried & Oswald Loretz. 1988. Die Keilalphabete. Die phonizisch kanaaniiischen und altarabischen Alphabete in Ugarit. ALASP 1. Munster.

-- . 1989. The Cuneiform Alphabets of Ugarit. Ugarit-Forschungen 21: 101-112. --. 1991. Die Keilalphabete aus Ugarit. In Cl. Baurain, C. Bonnet, V. Krings (eds. ) , Phoini-

keia Grammata. Lire et ecrire en Miditerranee. Actes du Colloque de Liege, 15-18 novembre 1989, pp. 49-67. Namur.

Folmer, Margaretha Louise. 1995. The Aramaic Language in the Achaemenid Period. A Study in Linguistic Variation. Orientalia Lovaniensia Analecta 68. Leuven.

Friedrich, Johannes & Wolf gang Rollig & Maria Giulia Amadasi Guzzo. 1999. PhonizischPunische Grammatik. 3. Auflage, neu bearbeitet von Maria Giulia Amadasi Guzzo unter Mitarbeit von Werner R. Mayer. Analecta Orientalia 55. Rome.

Freedman, David Noel. 1999. Psalm 119. The Exaltation of Torah (Biblical and Judaic Studies from the University of California, San Diego, v. 6), Winona Lake: Eisenbrauns 1999.

GaIT, W. Randall. 1985. Dialect Geography of Syria-Palestine, 1000-586 B. C.E. , Winona Lake: Eisenbrauns 1985.

Gelb, Ignaz J. 1963. A Study of Writing, Chicago. Gerhard, Volker. 2007. Die Einheit des Wissens: Aus Politik und Zeitgeschichte 46 /2007:

6-14· Getatchew Haile. 1996. Ethiopic Writing, in: Daniels & Bright (eds.) , The World's Writing

Systems, 1996 : 569-576. Greenfield, Jonas. 2001. Of Scribes, Scripts and Languages. Jonas C. Greenfield, Ai KarifCi

Yonah. Collected Essays 11, Leiden 2001, 173-185. Gruendler, Beatrice. 1993. The Development of the Arabic Scripts (Harvard Semitic Series

43) . Atlanta: Scholars Press 1993. --. 2006. Arabic Alphabet. Origin. Encyclopedia of Arabic Language and Linguistics,

Vol. I, Leiden 2006: 148-155. Gzella, Holger. (forthcoming). Abecedaries, in: Geoffrey Khan (ed.) , Encyclopedia of Hebrew

Language and Linguistics, Leiden (forthcoming).


Haider, Peter W. 1988-1989. Zu den agyptisch-agaischen Handelsbeziehungen zwischen ca. 1370 und 1200 v.Chr. : Miinstersche Beitriige zur Antiken Handelsgeschichte 7 (1988) 12-26; 8 (1989) 1-28.

Halayka, Issam KH. 2008. A Comparative Lexicon of Ugaritic and Canaanite (Alter Orient und Altes Testament 340) . Munster, Ugarit-Verlag 2008.

Hallo, .William W. 2004. Again the Abecedaries, in: Chaim Cohen, Avi Hurvitz, and Shalom M. Paul ( eds. ) , SeJer Moshe. The Moshe Weinfeldjubilee Volume. Studies in the Bible and the Ancient Near East, Qumran, and Post-Biblicaljudaism, Winona Lake, Indiana: Eisenbrauns 2004: 285-302.

Hawley, Robert. 2008. Apprendre a ecrire a Ougarit: une typologie des abecedaires, in: D'Ougarit it. jerusalem. Recueil d'etudes epigraphiques et arcMologiques offerts it. Pierre Bordreuil, M. Carole Roche (Orient & Mediterranee 2) . Paris: De Boccard 2008. 215-232.

Heide, Martin. 2007. Impressions from a New Alphabetic Ostracon in the Context of (Un) provenanced Inscriptions: Idiosyncrasy of a Genius Forger or a Master Scribe? In Meir Lubetski (ed. ) , New Seals and Inscriptions, Hebrew, Idumean, and Cuneiform, pp. 148-182. Sheffield Phoenic Press.

Hoch, James E. 1994. Semitic Words in Egyptian Texts of the New Kingdom and Third Intermediate Period. Princeton, New Jersey: Princeton University Press.

Irvine, A.K & A.F.L. Beeston. 1988. New evidence on the Qatabanian letter order. PSAS 18: 35-38.

Jeffery, L.H. 1961. The Local Scripts of Archaic Greece. A Study of the Origin of the Greek Alphabet and its Development from the Eighth to the Fifth Centuries B.C. . Oxford: Clarendon Press.

Kerr, Robert M. 2010. Von der aramaischen Lesekultur zur arabischen Schreibkultur. Kann die semitische Epigraphik etwas uber die Entstehung des Korans erzahlen? In Markus Gro� & Karl-Heinz Ohlig (eds.) , Die Entstehung einer Weltreligion L Von der koranischen Bewegung zum FrUhislam, pp. 354-376. Inarah. Schriften zur friihen Islamgeschichte und zum Koran, Band 5.

Knauf, Emst Axel. 1990. War "Biblisch-hebraisch" eine Sprache? Empirische Gesichtspunkte zur linguistischen Annaherung an die Sprache der althebrmschen Literatur. ZAH 3: 11-23.

Kooij, Gerrit van der. 1986. Early North-West Semitic Script Traditions. An Archaeological Study of the Linear Alphabetic Scripts upto c. 500 B.C. Dissertation, Leiden University.

Krahmalkov, Charles R. 2001. A Phoenician-Punic Grammar (HdO 54). Leiden: Brill. Krebemik, Manfred. 2007. Buchstabennamen, Lautwerte und Alphabetgeschichte. In Robert

Rollinger, Andreas Luther & Josef Wiesehofer (eds. ) , Getrennte Wege? Kommunikation, Raum und Wahmehmung in der Alten Welt, pp. 108-175. Frankfurt am Main: Verlag Antike.

Lehmann, Reinhard G. 2oo5a. Die Inschrift(en) des A/:urom-Sarkophags und die Schachtinschrift des Grabes V in jbeil (Byblos). Forschungen zur phonizisch-punischen und zyprischen Plastik lb. Dynastensarkophage mit szenischen Reliefs aus Byblos und Zypem Teil l.2. Mainz: Zabem.

--. 2oo5b. Space-Syntax and Metre in the Inscription ofYaQ.awmilk, King of Byblos, in: Omar Al-Ghul / A. Ziyadeh (ed.), Proceedings ofYarmouk Second Annual Colloquium on Epigraphy and Ancient Writings, Irbid. October 7th-9th, 2003, Faculty of Archaeology and Anthropology Publications 4, Yarmouk University, Irbid 2005, pp. 69-98.

--. 2008a. 'Who needs Phoenician?" Vom Nutzen des Phonizischen fur das Verstandnis der Sprache des Antiken Israel. Uberlegungen und Beispiele, in: Markus Witte / J.F. Diehl (Hg.) , Israeliten und PhOnizier - Ihre Beziehungen im Spiegel der Archiiologie und der Literatur des Alten Testaments, Fribourg / Gottingen 2008 (Orbis Biblicus et Orientalis 235) , S. 1-37.

--. 2008b. Calligraphy and Craftsmanship in the Ahirom Inscription: Considerations on Skilled Linear Flat Writing in Early 1st Millennium Byblos: Maarav 15.2 (2008) 119-164.


Lemaire, Andre. 1978. Abecedaires et exercices d'ecolier en epigraphie nord-ouest semitique:Journal Asiatique 266 (1978) 221-235.

--. 2008. "The Spread of Alphabetic Scripts (c. 1700-500 BCE)" : Diogenes 218 (2008) 45-58.

L[ 0 ] undin [ e] , A.G. 1987a. L'abecedaire de Beth Shemesh: Le Museon 100 (1987) 243-250. Lundin, A.G. 1987b. Ugaritic Writing and the Origin of the Semitic Consonantal Alphabet:

Aula Orientalis 5(1987) 91-99. Macdonald, Michael C.A. 1986. ABCs and Letter Order in Ancient North Arabian: Proceed

ings of the Seminar for Arabian Studies 16, Oxford 1986, 101-168. --. 2008a. Ancient North Arabian, in Roger D. Woodward ( ed. ) , The Ancient Languages

of Syria-Palestine and Arabia, Cambridge: University Press 2008: 179-224. --. 2008b. "The Phoenix of Phoinikeia. Alphabetic reincarnation in Arabia". In J. Baines,

J. Bennet and S. Houston (eds), The Disappearance of Writing Systems: Perspectives on Literacy and Communication. London: Equinox, 2008, pp. 207-229.

Martin, M.F. 1962. A Twelfth Century Bronze Palimpsest. Rivista degli Studi Orientali 37 (1962) 175-193.

Millard, Alan. 2007. Alphabetic Writing, Cuneiform and Linear, Reconsidered. Maarav 14(2) : 83-93.

-_ . 2009. Geschichte der Alphabetschrift. In: Holger Gzella (ed. ) , Sprachen aus der Welt des Alten Testaments, pp. 13-27. Darmstadt: WBG.

Misgav, Haggai & Yosef Garfinkel & Saar Ganor. 2009. The Ostracon. In Yosef Garfinkel and Saar Ganor (eds. ) , Khirbet Qetyafa Vol. 1. Excavation Report 2007-2008, pp. 243-257. Jerusalem: Israel Exploration Society.

Monroe, Christopher Mountfort. 2009. Scales of Fate. Trade, Tradition and Traniformation in the Eastern Mediterranenan ca. 1350-1175 BeE (AOAT 357) . Munster: Ugarit-Verlag.

Muller, Waiter W. 1994. Die aItsudarabische Schrift. In Hartmut Gunther & Otto Ludwig (eds.), Schrift und Schriftlichkeit. Writing and Its Use. Handbucher zur Sprach- und Kommunikationswissenschaft 10.1, pp. 307-312. Berlin, New York.

Naveh, Joseph. 1987. The Early History of the Alphabet. Jerusalem: Magnes. Nebes, Norbert & Peter Stein. 2008. Ancient South Arabian. In Roger D. Woodward (ed. ) ,

The Ancient Languages oJSyria-Palestine and Arabia, pp. 145-178. Cambridge: University Press.

Negbi, Ora. 1992. Early Phoenician Presence in the Mediterranean Islands: A Reapprisal. AmericanJournal of Archaeology 96: 599-615.

O'Connor, M. 1996a. Epigraphic Semitic Scripts. In: Peter T. Daniels & William Bright (eds.) , The World's Writing Systems, pp. 88-107. Oxford University Press.

--. 1996b. The Alphabet as Technology. In: Peter T. Daniels & William Bright (eds. ) , The World's Writing Systems, pp. 787-794. Oxford University Press.

Pardee, Dennis. 2003/2004. Review of: Josef Tropper, Ugaritische Grammatik, Munster 2000. Archiv for Orientforschung 50 (2203/2004), online version.

--. 2007. The Ugaritic Alphabetic Cuneiform Writing System in the Context of other Alphabetic Systems. In Cynthia 1. Miller ( ed.), Studies in Semitic and Afroasiatic Linguistics Presented to Gene B. Gragg, pp. 181-200. Chicago.

--. 2008. Ugaritic. In Roger D. Woodward (ed.) , The Ancient Languages of Syria-Palestine and Arabia, pp. 5-35. Cambridge University Press.

--. 2009. A New Aramaic Inscription from Zincirli. BASOR 356: 51-71. Powell, Barry P. 1991. Homer and the Origin oJ the Greek Alphabet. Cambridge University

Press. --. 2009. Writing. Theory and History of the Technology oJCivilization. Wiley-Blackwell. Puech, Emile. 1986. Origine de l'alphabet. Documents en alphabet lineaire et cuneiforme

du lIe millenaire. Revue Biblique 93: 161-213. -- . 1991. La tablette cuneiforme de beth She mesh. Premier temoin de la sequence des

lettres du sud-semitique. In Cl. Baurain, C. Bonnet, V. Krings (eds. ), Phoinikeia Grammata. Lire et ecrire en Mtiditerranee. Actes du eolloque de Liege, 15-18 novembre 1989, pp. 33-47. Namur.

Quack, Joachim Friedrich. 1993. Agyptisches und Sudarabisches Alphabet. Revue d'Egyptologie 44: 141-151.

--. 2003. Die spiitiigyptische Alphabetreihenfolge und das 'sudsemitische' Alphabet. LingAeg 11: 163-184.

Rainey, Anson F. 2007. Redefining Hebrew-A Transjordanian Language. Maarav 14(2) : 67-81.

Renz, Thomas. 2009. A Perfectly Broken Acrostic in Nahum l?joumal of Hebrew Scriptures 9(23) . [http:/ /purl.org/jhs ] .

Robin, Christian Julien. 2008. La lecture e t l'interpretation de l'abecedaire Ra's Shamra 88.2215. La preuve par l'Arabie? In Carole Roche (ed.), D'Ougarit a jerusalem. Recueil d'etudes epigraphiques et archeologiques offert a Pierre Bordreuil, pp. 233-244. Orient & Mediterranee 2. Paris: De Boccard.

Rollig, Wolfgang 1998. Nordsemitisch-Sudsemitisch? Zur Geschichte des Alphabets im 2. Jt.v.Chr. Israel Oriental Studies 18: 79-88.

Rollston, Christopher. 2008a. The Dating of the Early Royal Byblian Phoenician Inscriptions: A Response to Benjamin Sass. Maarav 15(1) : 57-93.

--. 2008b. The Phoenician Script of the Tel Zayit Abecedary and Putative Evidence for Israelite Literacy. In Ron Tappy & P. Kyle McCarter (eds. ) , Literate Culture and TenthCentury Canaan: The Tel Zayit Abecedary in Context, pp. 61-96. Eisenbrauns.

Ryckmans, J. 1985. L'ordre alphabetique sud-semitique et ses origines. In Ch. Robin (ed. ) , Melanges linguistiques offerts a Maxime Rodinson par ses eleves, ses collegues et ses amis, pp. 343-359. Paris.

Sanders, Seth L. 2008. Writing and early Iron Age Israel: Before National Scripts, Beyond Nations and States. In Ron Tappy & P. Kyle McCarter (eds.) , Literate Culture and TenthCentury Canaan: The Tel Zayit Abecedary in Context, 97-112. Eisenbrauns.

Saur, Markus. 2008. Der TyroszykLus des Ezechielbuches (BZAW 386) . Berlin: De Gruyter. Sass, Benjamin. 2004-2005. The Genesis of the Alphabet and its Development in the Sec

ond Millennium B. C. Twenty Years Later. KEN 2: 147-166. Schami, Rafik. 2008. Das Geheimnis des Kalligraphen. Munchen: Carl Hanser Verlag. Schipper, Bernd U. 2005. Die Erzahlung des Wenamun. Ein Literaturwerk im Spannungs

fold von Politik, Geschichte und Religion. Orbis Biblicus et Orientalis 209. Fribourg, Gottingen.

Schneider, Thomas. 1992. Asiatische Personennamen in agyptischen QueLLen des Neuen Reiches. Orbis Biblicus et Orientalis 114. Fribourg, Gottingen.

Schmitz, Philip C. 2011 (in press). Phoenician PRMN 'Pomegranate Fruit': An Unrecognized Compound Noun. Vicino Oriente 16.

Schulz, Hermann. 1987. Leviten im vorstaatlichen Israel und im Mittleren Osten. Munchen. Sherratt, Sus an. 1998. 'Sea Peoples' and the economic structure of the late second millen

nium in the eastern Mediterranean. In S. Gitin, A. Mazar, E. Stem (eds.) , Mediterranean Peoples in Transition. Thirteenth to Early Tenth Centuries BCE, pp. 292-313. Jerusalem, Israel Exploration SOciety.

Sherratt, Susan & Andrew. 1993. The Growth of the Mediterranean Economy in the Early First Millennium BC. World Archaeology 24(3) : 361-378.

Sivan, Daniel. 2001. A Grammar of the Ugaritic Language. HdO I, 28. Leiden: Brill. Sivan, Daniel & Cochavi-Rainey, Zipora. 1992. West-Semitic Vocabulary in Egyptian Script

of the 14th to the 10th Centuries BCE. Beer-Sheva. Studies by the Department of Bible and Ancient near East VI. Jerusalem: Ben-Gurion University of the Negev Press.

Smith, Duane. 2005. The Cuneiform Short Alphabet: Part 6. An Inscribed Jar handle from Sarepta in Lebanon. KTU 6(70) (Sar 3102, Pritchard (1975), 97-104) : <http://www.telecomtally.com/blog/ugarit/ KTU6. 70. pdf> .

Starcky, Jean. 1982. La fleche de Zakarba'al roi d'Amurru. In ArcheoLogie au Levant. Recueil a la memoire de Roger Saidah, pp. 179-186. Collection de la maison de I'Orient mediterraneen no 12 / Serie archeologique, 9. Lyon.

Stein, Peter. 2003. Untersuchungen zur Phonologie und Morphologie des Sabaischen. Epigraphische Forschungen auf der Arabischen Halbinse1 3. Rahden/Westf.: Leidorf.


--. 2010. Die altsiidarabischen Minuskelinschriften auf Holzstiibchen aus der Bayerischen Staastsbibliothek in Miinchen. Band 1: Die Inschriften der mittel- und spiitsabiiischen Periode. EFAH 5. Tubingen/Berlin: Wasmuth.

Steiner, Richard C. 2005. On the Dating of Hebrew Sound Changes (* If > If and *G > ' ) and Greek Translations (2 Esdras and Judith) .Joumal of Biblical Literature 124: 229-267.

Swiggers, Pierre. 1984. On the Nature of West-Semitic Writing Systems: Aula Grientalis 2: 149-15l.

--. 1991. Linguistic Considerations on Phoenician Orthography. In Cl. Baurain, C. Bonnet, V. Krings (eds.) , Phoinikeia Grammata. Lire et ecrire en Mfiditerranee. Actes du Co/toque de Liege, 15-18 novembre 7989, pp. 115-132. Namur.

Tappy, Ron E. & P. Kyle McCarter. (eds. ) 2008. Literate Culture and Tenth-Century Canaan: The Tel Zayit Abecedary in Context. Winona Lake: Eisenbrauns.

Tappy, Ron E. & P. Kyle McCarter & Marilyn J. Lundberg & Bruce Zuckerman. 2006. An Abecedary of the Mid-Tenth Century B.c.E. from the Judaean Shephela. BASGR 344: 5-46.

Tropper, Josef. 1994. Die nordwestsemitischen Schriften. In Hartmut Gunther & Otto Ludwig (eds. ) , Schrift und Schriftlichkeit. Writing and Its Use. Handbucher zur Sprachund Kommunikationswissenschaft 10.1, pp. 297-306. Berlin, N ew York.

-- . 2000. Ugaritische Grammatik. Alter Orient und Altes Testament 273. Munster: UgaritVerlag.

Ullendorff, Edward. 1977. Is Biblical Hebrew a Language? In Edward Ullendorff, Is Biblical Hebrew a Language? Studies in Semitic Languages and Civilizations, pp. 3-17. Wiesbaden.

Vennemann, Theo. 2006. Germanische Runen und phonizisches Alphabet. Sprachwissenschaft 34: 367-429.

-- . 2009. Zur Reihung der Runen im aIteren Fupark. Analecta Septentrionalia-RGA-E 65: 832-86l.

Vita, Juan-Pablo. 2003. La protohistoria de la lengua fenicia. In A. GonzaJez Blanco, J.P. Vita, J.A. Zamora (eds. ) , De la Tablilla a la Inteligencia Artificial I, pp. 395-410. Zaragoza.

Wevers, J.W. 1970. Ifeth in Classical Hebrew. In J.W. Wevers, D.B.Redford (eds. ) , Essays on the Ancient Semitic World, pp. 101-112. Toronto.

Walde, A. & J.B. Hofmann. 1938. Lateinisches Etymologisches Worterbuch. 3., neubearbeitete Auflage. Heidelberg.

Wachter, Rudolf. 1989. Zur Vorgeschichte des griechischen Alphabets. Kadmos 28: 19-78. -- . 1996. Alphabet. In Der neue Pauly. Enzyklopiidie der Antike, pp. 536-548. Stuttgart:

Metzler. Whittaker, Molly. 1982. Tatian Gratio ad Graecos and fragments. Edited and translated.

Oxford: Clarendon Press. Young, lan. 1993. Diversity in Pre-Exilic Hebrew (FAT 5). Tubingen: Mohr.

Writing Accross Boarders

Documents

Transcript of Writing Accross Boarders